Yes Africa Can

Success Stories from a Dynamic Continent

Punam Chuhan-Pole
Manka Angwafo
Editors
Yes Africa Can
Yes Africa Can
SUCCESS STORIES FROM A DYNAMIC
CONTINENT
EDITORS
PUNAM CHUHAN-POLE AND MANKA ANGWAFO
© 2011 The International Bank for Reconstruction and Development / The World Bank
1818 H Street NW
Washington DC 20433
Telephone: 202-473-1000
Internet: www.worldbank.org

All rights reserved

1 2 3 4 14 13 12 11

This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The
findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive
Directors of The World Bank or the governments they represent.
The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denomina-
tions, and other information shown on any map in this work do not imply any judgement on the part of The World Bank
concerning the legal status of any territory or the endorsement or acceptance of such boundaries.

Rights and Permissions

The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permis-
sion may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank
encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly.
For permission to photocopy or reprint any part of this work, please send a request with complete information to the
Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470;
Internet: www.copyright.com.
All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher,
The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org.

ISBN: 978-0-8213-8745-0
eISBN: 978-0-8213-8746-7
DOI: 10.1596/978-0-8213-8745-0

Cover design: Naylor Design, Inc.

Library of Congress Cataloging-in-Publication Data

Yes Africa can : success stories from a dynamic continent / editors: Punam Chuhan-Pole, Manka Angwafo.
p. cm.
Includes bibliographical references.
ISBN 978-0-8213-8745-0 — ISBN 978-0-8213-8746-7 (electronic)
1. Africa—Economic policy. 2. Africa—Economic conditions—1960- 3. Africa—Economic conditions—
21st century. 4. Africa—Social conditions. 5. Africa—Politics and government—21st century. I. Chuhan-Pole, Punam.
II. Angwafo, Manka.

HC800.Y47 2011
330.967—dc22 2011014918
 CONTENTS

Foreword ix
Acknowledgments xi
Abbreviations xiii

Overview 1
Punam Chuhan-Pole and Shantayanan Devarajan

PART I
TRACKING SUCCESSFUL GROWTH EXPERIENCES
Chapter 1 Tanzania: Growth Acceleration and Increased Public Spending with
Macroeconomic Stability 21
David O. Robinson, Matthew Gaertner, and Chris Papageorgiou

Chapter 2 Building on Growth in Uganda 51

Sarah Ssewanyana, John Mary Matovu, and Evarist Twimukye

Chapter 3 Rapid Growth and Economic Transformation in Mozambique, 1993–2009 65

Antonio M. D. Nucifora and Luiz A. Pereira da Silva

Chapter 4 Botswana’s Success: Good Governance, Good Policies, and Good Luck 81
Michael Lewin

Chapter 5 Mauritius: An Economic Success Story 91

Ali Zafar

Chapter 6 Cotton Dependence in Burkina Faso: Constraints and Opportunities for Balanced Growth 107
Jonathan Kaminski

PART II
POST CONFLICT SITUATIONS – BUILDING INSTITUTIONS AND GOVERNANCE
Chapter 7 Postconflict Economic Governance Reform: The Experience of Liberia 127
Vishal Gujadhur

Chapter 8 Decentralization in Postconflict Sierra Leone: The Genie Is Out of the Bottle 141
Vivek Srivastava and Marco Larizza

Chapter 9 Transport Infrastructure and the Road to Statehood in Somaliland 155

Jean-Paul Azam
v
 PART III
LEVERAGING SECTORAL ADVANTAGES TO EXPAND EXPORTS
Chapter 10 Growing Mali’s Mango Exports: Linking Farmers to Market through
Innovations in the Value Chain 167
Yéyandé Sangho, Patrick Labaste, and Christophe Ravry

Chapter 11 Economic Liberalization in Rwanda’s Coffee Sector: A Better Brew for Success 185
Karol C. Boudreaux

Chapter 12 Cocoa in Ghana: Shaping the Success of an Economy 201

Shashi Kolavalli and Marcella Vigneri

Chapter 13 Apparel Exports in Lesotho: The State’s Role in Building

Critical Mass for Competitiveness 219
Mallika Shakya

Chapter 14 The Success of Tourism in Rwanda: Gorillas and More 231

Hannah Nielsen and Anna Spenceley

PART IV
BOOSTING AGRICULTURAL EFFICIENCY AND OUTPUT THROUGH TARGETED INTERVENTIONS
Chapter 15 Increasing Rice Productivity and Strengthening Food Security through
New Rice for Africa (NERICA) 253
Aliou Diagne, Soul-Kifouly Gnonna Midingoyi, Marco Wopereis, and Inoussa Akintayo

Chapter 16 Fertilizer in Kenya: Factors Driving the Increase in Usage by Smallholder Farmers 269
Joshua Ariga and T. S. Jayne

Chapter 17 Malawi’s Agricultural Input Subsidy Program Experience over 2005–09 289
Andrew Dorward, Ephraim Chirwa, and T. S. Jayne

Chapter 18 MoneyMaker Pumps: Creating Wealth in Sub-Saharan Africa 319

I. V. Sijali and M. G. Mwago

PART V
ENGAGING THE PRIVATE SECTOR TO UPGRADE INFRASTRUCTURE
Chapter 19 ICT in Sub-Saharan Africa: Success Stories 339
Kaoru Kimura, Duncan Wambogo Omole, and Mark Williams

Chapter 20 Mobile Payments Go Viral M-PESA in Kenya 353

Ignacio Mas and Dan Radcliffe

Chapter 21 Independent Power Projects in Sub-Saharan Africa: Determinants of Success 371

Anton Eberhard and Katharine Nawal Gratwick

PART VI
IMPROVING HUMAN DEVELOPMENT OUTCOMES WITH INNOVATIVE POLICIES
Chapter 22 Innovative Financing for Health in Rwanda: A Report of Successful Reforms 403
C. Sekabaraga, A. Soucat, F. Diop, and G. Martin

Chapter 23 The Malaria Control Success Story 417

Anne-Maryse Pierre-Louis, Jumana Qamruddin, Isabel Espinosa, and Shilpa Challa

Chapter 24 Health Extension Workers in Ethiopia: Improved Access and

Coverage for the Rural Poor 433
Nejmudin Kedir Bilal, Christopher H. Herbst, Feng Zhao, Agnes Soucat, and Christophe Lemiere

vi CONTENTS
Chapter 25 Family Planning Trends in Sub-Saharan Africa: Progress, Prospects,
and Lessons Learned 445
Mona Sharan, Saifuddin Ahmed, John May, and Agnes Soucat

Chapter 26 Achieving Universal Primary Education through School Fee Abolition:

Some Policy Lessons from Uganda 465
B. Essama-Nssah

CONTENTS vii
 F O R E WO R D
In April 2011, a study entitled “Hiding the Real Africa” doc-
umented how easily Africa makes the evening newscasts
and newspaper headlines in the West when a major famine,
pandemic, or violent crisis breaks. For example, over a five-
month period from May 2010, more than 245 articles on
Africa published by the 10 most-read US newspapers
focused on poverty. Only five of them mentioned wealth
and growth.
The tendency to dwell on Africa’s challenges, long-
standing problems, and failures rather than on its opportu-
nities and successes is one of the continent’s most enduring
stereotypes. Despite an acceleration of economic growth
over the past 10 years and a growing army of African mid-
dle-class consumers, the narrative about Africa has
remained one of poverty, disease, and conflict.
This suggests that changing the narrative on Africa
requires not only sustained economic progress in the con-
tinent, but also a collaborative platform for telling the story
in a compelling way.
Economic growth has spurted, averaging 5 percent a year
from 1998 to 2008. Despite a sharp slowdown in the wake of
the recent global financial crisis, growth has rebounded and
is back on track. Africa’s private sector is showing signs of
dynamism, attracting growing amounts of investment from
domestic and foreign investors. Africa’s poverty rate declined
by about 1 percentage point a year, from 59 percent in 1995
to 51 percent in 2005. Primary completion rates are rising
faster in Africa than anywhere else. More girls are in school,
as educational opportunities for girls have expanded across
the region. The death of children aged five or younger con-
tinues to decline.
Undeniably, much remains to be done: nearly 400 mil-
lion people still live in extreme poverty; human capital is
low; the production structure of African economies is
largely undiversified, heavily concentrated in primary com-
modities; and governance remains weak.
To harness the recent growth and dynamism in the con-
tinent to address these long-term challenges, we need to
understand what was behind Africa’s resurgence. This
book documents some of the success stories in Africa. The
26 case studies, which have been prepared by local and
international academics, analysts, and practitioners, take
an in-depth look at economic and social development
achievements across countries, themes, and sectors. Some
of these are well known, such as Mauritius’s economic
growth experience and the information, communications,
and technology revolution in Africa. But others are not:
the transformation of Rwanda’s coffee sector and the shift
to high-quality coffee production; Lesotho’s success in
boosting apparel exports; Mozambique’s rapid economic
growth (averaging more than 8 percent a year from
1993–2009); and the progress in combating malaria. In
addition to documenting the nature of the successes, the
case studies identify the reasons and draw lessons for other
countries on the continent and elsewhere.
I expect Yes Africa Can: Success Stories from a Dynamic
Continent to stimulate the interest of both the media and
filmmakers to seek out and tell more of Africa’s many success
stories; to help put Africa’s positive achievements on the map;
and to inspire a vigorous discussion on how Africans can do
more to unleash the full economic potential that is needed to
transform one of the world’s fastest growing regions.
Obiageli Katryn Ezekwesili
Vice President, Africa Region
The World Bank
ix
 AC K N OW L E D G M E N T S
This volume was prepared by a team led by Punam
Chuhan-Pole under the general guidance and direction of
Shantayanan Devarajan, Chief Economist of the World
Bank’s Africa Region. Several of the case studies included in
this volume were written by local researchers and practi-
tioners based in Sub-Saharan Africa.
The team contributing to the book consisted of:
overview—Punam Chuhan-Pole, Shantayanan Devarajan,
Manka Angwafo and Dana Vorisek; chapter 1—David
Robinson, Matthew Gaertner, and Chris Papageorgiou; chap-
ter 2—Sarah Ssewanyana, John May Matovu, and Evarist
Twimukye; chapter 3—Antonio Nucifora and Luiz A. Pereira
da Silva; chapter 4—Michael Lewin; chapter 5—Ali Zafar;
chapter 6—Jonathan Kaminski; chapter 7—Vishal Gujadhur;
chapter 8—Vivek Srivastava and Marco Larizza; chapter 9—
Jean-Paul Azam; chapter 10—Yéyandé Sangho, Patrick
Labaste, and Christophe Ravry; chapter 11—Karol Boudreaux;
chapter 12—Shashi Kolavalli and Marcella Vigneri; chapter
13—Mallika Shakya; chapter 14—Hannah Nielson and Anna
Spenceley; chapter 15—Aliou Diagne, Soul-Kifouly Gnonna
Midingoyi, Marco Wopereis, and Inoussa Akintayo; chap-
ter 16—Joshua Ariga and T.S. Jayne; chapter 17—Andrew
Dorward, Ephraim Chirwa, and T. S. Jayne; chapter 18—
I.V. Sijali and M. G. Mwago; chapter 19—Kaoru Kimura,
Duncan Wambogo Omole, and Mark Williams; chapter 20—
Ignacio Mas and Dan Radcliffe; chapter 21—Anton Eberhard
and Katharine Nawal Gratwick; chapter 22—C. Sekabaraga,
A. Soucat, F. Diop, and G. Martin; chapter 23—Anne-Maryse
Pierre-Louis, Jumana Qamruddin, Isabel Espinosa, and Shilpa
Challa; chapter 24—Nejmudin Kedir Bilal, Christopher
H. Herbst, Feng Zhao, Agnes Soucat, and Christophe Lemiere;
chapter 25—Mona Sharan, Saifuddin Ahmed, John May, and
Agnes Soucat; and chapter 26—B. Essama-Nssah.
Many people provided guidance in identifying potential
success stories. Notably, Cecilia M. Briceno-Garmendia,
Donald F. Larson, Hannah Messerli, Vincent Palmade,
Markus Moeller, Agnes Soucat, and Hassan Zaman. Discus-
sions with Jorge Saba Arbache, Christopher Paul Jackson,
and Stephen Mink were also useful in selecting case studies.
Laurent Wagner made a central contribution in the early
phase of the study.
The report benefited from input by peer reviewers: Rocio
Castro, Vandana Chandra, Quy-Toan Do, Elena Ianchovichina,
Hannah Messerli, Celestin Monga, Blanca Moreno-Dodson,
Claudia Paz Sepulveda and Hassan Zaman. Comments were
also received from Fabiano Bastos, Jaime Biderman, Maya
Brahmam, Mukesh Chawla, Delfin Go, Monica Das Gupta,
Kenechukwu Maria Ezemenari, Errol George Graham, Jane
Wangui Kiringai, Patrick Labaste, Karen Mcconnell Brooks,
Moussa Diarra, Motoky Hayakawa, Maureen Lewis, Christine
Lao Pena, Anne M. Pierre-Louis, Kofi Nouve, Rojid Sawkut,
Shahzad Sharjeel, Toni Sittoni, Yutaka Yoshino, and Ali Zafar.
Participants at the April 2010 forum provided useful
insights and perceptives that helped shape the output. Excel-
lent editorial review was provided by Dana Vorisek, Barbara
Karni, and Martha Gottron. Ann G. Karasanyi and Kenneth
Omondi provided administrative support throughout the
project, particularly during the forum. The Web site featuring
the stories was produced by Michael Matovina and Yohannes
Kebede. Mapi M. Buitano and Odilia Renata Hegba managed
the communication and dissemination activities. Book
design, editing, production, and printing were coordinated by
Steven McGroarty, Rick Ludwick, and Denise Bergeron of the
World Bank Office of the Publisher.
Finally, the production of this volume was made possible
by the World Bank Office of the Publisher.
Punam Chuhan-Pole
Manka Angwafo
Editors
xi
 CH
A BB
ARPETVEIRAT
T IHORN
E SE

ACT artemisinin-based combination therapy

BPO business process outsourcing
CPR contraceptive prevalence rate
DHS Demographic and Health Survey
ECOWAS Economic Community of West African States
FAO Food and Agriculture Organization
FMOH Federal Ministry of Health
GDP gross domestic product
GSM Global System for Mobile Communications
HEP Health Extension Program
ICT information and communications technology
IDA International Development Association
IT/ITES information technology/information technology–enabled service
MOES Ministry of Education and Sports
NERICA New Rice for Africa
NGO nongovernmental organization
PC personal computer
PIN personal identification number
PMA Plan for the Modernization of Agriculture
SADC Southern Africa Development Community
SMS short messaging system
TFR total fertility rate
UNFPA United Nations Population Fund
USAID U.S. Agency for International Development
WAEMU West African Economic and Monetary Union
WHO World Health Organization
(all dollar figures are U.S. dollars)

xiii
 Overview
Punam Chuhan-Pole and Shantayanan Devarajan
WHY STUDY AFRICAN SUCCESSES?
Over the past decade Sub-Saharan Africa has seen a remark-
able turnaround in economic performance. After years of
stagnation, economic growth has spurted—gross domestic
product (GDP) grew from an annual average rate of less
than 2 percent in 1978–95 to nearly 6 percent over 2003–08.
Inflation is half its level of the mid-1990s. Private capital
flows have risen to $50 billion, exceeding foreign aid. Exports
are growing, as is private sector activity. The number of dem-
ocratic regimes has risen and the security situation has
improved. The poverty rate is falling by 1 percentage point a
year. Countries such as Ethiopia, Ghana, Mauritania, and
Rwanda are on track to reach many of the Millennium
Development Goals. Nine African countries have
achieved or are on track to achieve the target for extreme
poverty (World Bank and IMF 2011). Among other encour-
aging trends are more fair and effective leadership, an
improving business climate, increasing innovation, a more
involved citizenry, and growing reliance on home-grown
solutions. More and more, Africans are driving African
development.
This increased dynamism in Sub-Saharan Africa is evi-
dent across a broad swath of countries. It has created opti-
mism that Africa’s favorable development performance will
be long lasting and that it could dramatically transform
countries in the region. Along the way, the prevailing
We wish to thank Manka Angwafo and Dana Vorisek for their excellent contributions.
discourse on Africa’s economic development has shifted
from whether the region will develop to how the region is
developing.
Yet, there are still causes for concern. For one thing, per-
formance across countries varies substantially. Also worry-
ing is that, historically, Africa’s performance has been
volatile—with short periods of acceleration followed by
long periods of deceleration and decline brought on by
recurrent crises (Arbache and Page 2007). Moreover, the
pattern of progress underscores serious shortfalls in some
areas, notably, in the economic diversification of many
countries and in the integration of African economies into
the global economy (UNIDO 2009). African countries also
need to correct large infrastructure deficits, dramatically
expand the skills base of their labor forces, and improve
their ability to absorb technical knowledge in the private
sector to improve their global standing in the years ahead.
With the lingering specter of past failures, continued
economic development challenges in Sub-Saharan Africa
have led to questions of whether improved performance in
recent years is sustainable. This book attempts to answer
these questions by documenting and better understanding
some of the impressive achievements that have occurred in
recent years. By systematically identifying and assessing pos-
itive outcomes, it is possible to draw out lessons regarding
what has worked in practice and why. These lessons will in
1
 turn inform our judgment about whether Africa’s growth is
sustainable.
After a review of the major recent economic develop-
ments in Africa, this overview describes the approach and
methodology used in the study of African successes. We then
present a framework for understanding the 26 case studies, a
framework that is based on the balance between overcoming
market failures and overcoming government failures. We
conclude that the way in which African governments have
been increasingly addressing government failures—with
impressive results—bodes well for the sustainability of
Africa’s decade-long growth.
A CHANGING ECONOMIC LANDSCAPE
After lackluster economic performance for decades, during
which the gap between Sub-Saharan Africa and the rest of
the developing world widened, the region’s economies have
seen a visible turnaround that began in the mid-1990s. The
most visible evidence is the acceleration in output growth
from under 2 percent in 1978–95 to nearly 6 percent in
2003–08.1 Oil exporters have seen an especially steep rise,
but growth has been widespread (figure 1). Indeed, 21
non-oil countries, home to over 40 percent of the region’s
population and accounting for nearly a quarter of the
region’s GDP, have posted annual economic growth of
more than 4 percent during 1995–2008. Many of these
21 countries face the additional challenge of being land-
locked. Within this overall improvement in the growth
trend, some countries are lagging, but their numbers are
small and declining.
In a remarkable break from historical performance, not
only has growth accelerated, but the acceleration has been
sustained for a longer period than in the past. For example,
Mozambique, Tanzania, and Uganda have experienced
growth rates above 5 percent in every year during 2001–08.
Arbache et al. (2008) find that growth accelerations were
more frequent in 1995–2005, whereas growth decelerations
were more common in the preceding two decades: the like-
lihood of a growth acceleration was 0.42 in 1995–2005 but
0.21 in 1985–94.
Strong output growth has lifted per capita incomes as
well. Per capita incomes fell in 1975–84 at an average annual
rate of 0.88 percent, and at an even faster pace (1.13 percent)
in 1985–94. By contrast, per capita income expanded at
nearly 2 percent a year in 1995–2008, in step with growth
rates in other developing regions (figure 2).
African exports are growing as well but remain highly
concentrated (figure 3). After stagnating in volume and
2 OVERVIEW
value terms in the 1980s, exports doubled in volume and
rose fivefold in value in the period 1994–2008. Nevertheless,
exports of goods (excluding oil) account for less than
20 percent of GDP (reflecting little change over the previous
decades), and Africa’s share of world exports remains low,
after falling for more than half a century. Export growth has
been led by extractive resource exports, fueled by a com-
modity boom (figure 4). Africa’s exports continue to be
dominated by extractive and agricultural commodities. In
many cases, commodity dependence seems to have intensi-
fied. Manufacturing exports have grown slowly, and much
of the region has yet to break into the global industrial mar-
ket (UNIDO 2009). The diversity and sophistication of
exports has changed very little. Whereas the mostly upward
trend in extractive commodity prices has provided revenues
to finance much needed infrastructure and social spending,
it also presents challenges of avoiding the “resource curse”
and converting the revenues into sustained development.
Improving economic prospects have attracted foreign
direct investment (FDI) to the region (figure 5). FDI grew by
more than 17 percent a year between 1995 and 2008. Although
extractive industries account for the bulk of the FDI inflows,
significant amounts—especially greenfield investments—are
being recorded in the construction, communications, and
banking sectors. Emerging market countries such as China
and India are becoming important sources of FDI.
A notable feature of the transformation under way in
Africa is the upsurge in agricultural output and productiv-
ity. Agricultural GDP growth averaged nearly 3 percent a
year over 1980–2004, but growth per capita averaged only
0.9 percent, between one-third and half that of other devel-
oping regions (World Bank 2008). This average hides con-
siderable variation across countries and over time. For
example, per capita growth in agricultural output was neg-
ative in the 1980s and early 1990s, before turning positive. A
recent study (Fuglie 2010) finds that the pickup in agricul-
tural growth in the 1990s and 2000s resulted from expan-
sion of land under agriculture. Block (2010) finds that agri-
cultural total factor productivity growth declined in 1960s
and 1970s, but productivity picked up over the last two
decades. Signs of change in African agriculture are evident
in the successes documented in this book, such as the
expansion in maize production in Kenya that was driven by
liberalization of the local fertilizer market, in cotton in West
Africa, and in cassava and rice production in many coun-
tries, boosted by improved varieties. In a region where two-
thirds of the population is rural and 70 percent of the poor
derive their livelihood from agriculture, the favorable trend
in agriculture is encouraging. Robust growth in agriculture
Figure 1 Average Real GDP Growth Rates in Sub-Saharan Africa: 1995–2008

Equatorial Guinea
Angola
Chad
Sudan 31% of African
Nigeria population
Cameroon
Congo, Rep.
Gabon

Liberia
Rwanda
Mozambique
Uganda
Ethiopia
Sao Tome and Principe
Cape Verde
Burkina Faso
Botswana
Mali
Tanzania 41% of African
Ghana
population
Malawi
Benin
Sierra Leone
Gambia, The
Namibia
Mauritius
Mauritania
Senegal
Niger
Zambia
Madagascar

Guinea
South Africa
Togo
Kenya
Lesotho
Swaziland
Seychelles 27% of African
Cote d'Ivoire population
Comoros
Central African Republic
Congo, Dem. Rep.
Guinea-Bissau
Eritrea
Burundi
Zimbabwe
–5 0 5 10 15 20 25 30
percent

Source: World Development indicators, World Bank.

holds the promise of enhancing food security and facilitat-
ing broad-based economic growth and poverty reduction.
As stagnation in African economies has given way to
growth, poverty has begun to decline. The incidence of
extreme poverty—the share of people living on less than
$1.25 a day—fell from 59 percent in 1995 to nearly 50 per-
cent in 2005, a decline of 1 percentage point a year in the
poverty rate. Several African countries are likely to achieve
the MDG of halving income poverty by 2015 (Goal 1).
Overall, however, the region lags other regions in attaining
Goal 1 of the MDGs.
Education and health indicators have also improved
(figure 6). For example, primary school enrollment has
jumped 14 percentage points, from about 59 percent in

OVERVIEW 3
 Figure 2 Average per capita income growth by country Figure 4 Trends in Commodity Prices and Terms of
groups: 1975–2008 Trade in Sub-Saharan Africa

4.5 a. Commodity and oil prices

4.0 400 120

2000 = 100 using constant US$

3.5 350
3.0 100

$/bbl, constant 2000 US$

2.5 300
percent

2.0 80
1.5 250
1.0 200 60
0.5
0 150
–0.5 40
0 100
–1.5 20
50
a n

e
ric ra

m
0 0
Af Saha

co

co

co
in

in

in
b-

0
gh

–7

–8

–8

-9

–9

–0

–0

–1
dl
Su

lo
hi

Jan
id

Jan

Jan

Jan

Jan

Jan

Jan

Jan
m

1975–1984 1985–1994 1995–2008 metals and minerals agriculture crude oil brent, in $/bbl

Source: World Development indicators, World Bank. b. Sub-Saharan Africa: Terms-of-trade index
Note: Middle and low income includes Sub-Saharan Africa. 160

index, 2000 = 100

140

120
Figure 3 Trends in Export Volumes and Values,
Sub-Saharan Africa, 1970–2010 100

a. Export volumes
300 80
80

85

90

95

00

05

10
250
19

19

19

19

20

20

20
in constant US$ billions

200 Source: World Development indicators, World Bank.

150
2000 to 73 percent in 2008—the fastest improvement of any
100 region. More girls are attending school. Some significant
50 gains have also been made in the area of health, where
progress has admittedly been mixed. For example, child
0
deaths (under-five mortality rates) have declined from 181
70

74

78

82

86

90

94

98

02

06

10
19

19

19

19

19

19

19

19

20

20

20

per 1,000 births in 1990 to 132 per 1,000 in 2009, with some
b. Export values of the poorest countries, such as Eritrea and Malawi, show-
600
ing remarkable progress. Maternal mortality rates are also
500 trending down, though not fast enough, and the region is
in current US$ billions

400
beginning to make inroads in halting the spread of commu-
nicable diseases.
300 Since the mid-1990s African countries have made strides
200 in the area of economic management and structural poli-
cies. Combined with debt relief, these reforms have helped
100
to redress external and fiscal imbalances and rein in infla-
0 tion. Median inflation has fallen from double-digit levels in
70

74

78

82

86

90

94

98

02

06

10

the 1970s and 1980s to well below 10 percent in 1996–2007.

19

19

19

19

19

19

19

19

20

20

20

Source: World Development indicators, World Bank.

In 1995 nearly a third of countries had inflation rates above
20 percent; in 2008 this figure had dropped to one-tenth

4 OVERVIEW
Figure 5 Foreign Direct Investment in Sub-Saharan Africa

40 5

35 4.5
4
30
3.5

percent of GDP (%)

billions of US$

25 3
20 2.5

15 2
1.5
10
1
5 0.5
0 0
87

93

96

02
78

81

84

90

99

05

08
19

19

19

20
19

19

19

19

19

20

20
foreign direct investment, net inflows (BoP, current US$)
foreign direct investment, net inflows (% of GDP)

Source: World Development indicators, World Bank.

Figure 6 Progress in Human Development in (figure 7). Exchange rates have also been maintained at
Sub-Saharan Africa competitive levels, including the one-time devaluation of
the CFA franc in 1994.
a. Primary school enrollment rate Over the past several years, many African countries have
80 60
moved to liberalize trade, and reforms have brought down
70
40 tariff rates. Average most-favored-nation (MFN) applied
percent

millions

60 tariff rates more than halved from 1981 to 2009 and are
20 comparable to those of developing countries in other
50
regions (figure 8). But improvements in trade facilitation
40 0 have lagged. In a bid to enhance competitiveness, African
countries have also made progress in recent years in imple-
99

01

03

05

07
19

20

20

20

20

menting reforms to support the investment climate.

total primary-age children out of school
Because of reforms, Rwanda is now one of the fastest places
net primary enrollment rate
in the world to start a business (11th overall according to
the World Bank’s 2010 Doing Business report). There are
b. Under-five mortality rate
also indications that the quality of governance is improving
181 in some countries (figure 9). But, overall, weak governance
180
under-five mortality
rate (per 1000)

remains a problem in Africa, especially in fragile states and

155
resource-rich countries, and exerts a drag on long-term
130 132 development.
105

80 APPROACH AND METHODOLOGY

90

95

00

05

06

07

08

09
19

19

20

20

20

20

20

20

A number of major academic and policy studies pub-

Source: World Development indicators, World Bank. lished in recent years have examined trends related to

OVERVIEW 5
 Figure 7 Inflation Patterns in Sub-Saharan Africa economic development in Sub-Saharan African countries.
In Can Africa Claim the 21st Century?, released in 2000,
100 authors from four multilateral institutions conclude that
16.7 10.3
31.6 the emergence of three favorable trends—greater political
percent of countries

80 36.0
33.3 participation (the basis for more accountable govern-
48.7
60 ments), a changing view of Africa in the wake of the end
28.9
40 of the cold war, and the increasing presence of globaliza-
52.0
tion and information technology—hold the promise of
50.0 41.0
20 39.5 accelerating development and moving Africa out of
12.0 poverty. At the same time, the book indicates that
0
1975 1985 1995 2008 strengthening governance, investing in people, increasing
under 10% inflation 10–20% inflation competitiveness, and reducing aid dependence are crucial
above 20% inflation to development paths in African economies.
Source: World Development indicators, World Bank. Ten years later, Radelet, in his book Emerging Africa: How
Note: The figure shows the share of countries in each inflation range. 17 Countries Are Leading the Way, finds that an improving
governance environment, stronger economic policies, a
Figure 8 Trends in Average MFN Applied Tariff Rates, changed relationship with the international donor commu-
Sub-Saharan Africa, 1981–2008 nity, the increased accountability provided through new
technologies, and a new generation of business leaders and
35 policy makers have been indispensable in bringing about
percent, unweighted

30 progress in the 17 African countries where a turnaround has

25 been most evident.
20
Drawing on the existing knowledge of African develop-
15
ment presented in these and other publications,2 Yes Africa
10
5 Can: Success Stories from a Dynamic Continent takes an in-
0 depth look at 26 economic and social development successes
in Sub-Saharan African countries—20 from individual
5

8
–8

–9

–9

00

–0

–0
81

86

91

–2

01

06

countries and 6 that occurred regionwide.3 The successes

19

19

19

96

20

20
19

manifest at the project, provincial, subnational, national, or

Sub-Saharan Africa developing countries (134)
regional level and cut across themes, programs, and sectors.
Source: World Development indicators, World Bank. The overall goal of the research is to address how Sub-
Saharan African countries have overcome major develop-
ment challenges. The approach straddles the development
Figure 9 Trends in Political Rights and Civil Liberties in debate that Cohen and Easterly explore in What Works In
Sub-Saharan Africa Development?: Thinking Big and Thinking Small. That is,
should one adopt a big-picture approach that focuses on the
7
role of the quality of institutions and macroeconomic and
6
structural policies or an approach that focuses on micro-
5
economic interventions such as conditional cash transfers
4
and bed nets, whose impact can be measured through
3
impact evaluations?
2 The case studies were chosen through a stocktaking
1 exercise that included consultation with experts inside and
0 outside the World Bank. To qualify for selection, each case
1974 1979 1984 1989 1994 1999 2004 2009
study had to meet several criteria. First, the achievement
political rights civil liberties
identified in the case study had to be observable and meas-
Source: Freedom House. urable. Second, the achievement had to represent an outcome
Notes: Political rights and civil liberties are measured on a one-to-seven
scale, with one representing the highest degree of freedom and seven the perceived as desirable in the literature on development.
lowest. Third, there must have been existing analytical work—impact

6 OVERVIEW
evaluations, assessments (including against relevant regional
or sectoral benchmarks), and reports—regarding the
achievement. And fourth, the achievement must have been
sustained over time. In light of the above criteria, and
because of the potential lessons that can be learned, there
was a focus on critically evaluating well-recognized success
stories in addition to those that are less well known. Where
possible, the case studies attempt to identify achievements
that have the potential to be scaled up or that have elements
of transferability.
The coverage of development successes is selective rather
than comprehensive. Indeed, some achievements were not
examined because of resource constraints and timely avail-
ability of experts who could undertake the studies. A
notable example that was not studied in detail is the success
of Kenya’s horticulture industry in becoming a major sup-
plier of cut flowers to the European Union.
The 26 case studies presented in Yes Africa Can: Success
Stories for a Dynamic Continent cover six broad topical areas:
Successful growth experiences. Both a high rate of
growth and the quality of growth are central to achieving
poverty reduction.
■ Sustaining strong economic growth. High and sustained
economic growth is needed if countries are to substan-
tially reduce poverty and achieve the MDGs. A few African
countries have grown at a brisk pace over many years.
What explains these long periods of growth accelerations?
What has been the role of a strong business environment
in unleashing growth? Has growth been inclusive?
■ Quality of growth—industrial development. Empirical
evidence shows that countries with more diversified pro-
duction and more sophisticated production sectors have
stronger growth. Africa’s industrial development has
been disappointing. Yet, there are glimmers of progress.
How have some African countries achieved more broad-
based production?
Postconflict situations: building institutions and
Governance. Institutions and governance are central to the
development process anywhere, but they are particularly
vital in postconflict situations. How have some postconflict
countries established better governance and transparency?
Expanding exports. Increasing competitiveness and diver-
sifying exports is important if a country is to integrate in the
global economy and benefit from globalization. How have
some countries managed to diversify exports away from
traditional goods? How have some countries been able to
build capacity so as to respond to quantity and quality
needs of the global market, including compliance with
international standards?
Boosting agricultural efficiency and output through
targeted interventions and reforms. The vast majority
of Africa’s poor are concentrated in rural areas and are heavily
dependent upon agriculture for their livelihoods. Spurring
agricultural development will be vital for achieving the
MDGs in Africa. Raising agricultural productivity is impor-
tant for stimulating growth in other sectors as well. What are
some of the ways in which countries have succeeded in boost-
ing agricultural production, commercial agriculture, and
productivity?
Engaging the private sector to upgrade infrastruc-
ture. Africa has large unmet infrastructure needs: the latest
estimates point to a financing gap of more than $30 billion
a year (World Bank 2010). Closing the infrastructure gap
will require more private investment and operation. But the
region has been slow to mobilize private resources, partly
because of governance, institutional, and regulatory issues.
How have some countries managed to attract private
participation in delivering infrastructure services? One
important area where African countries have seen rapid
progress is in the mobile telecommunications segment of
the information and communications technology (ICT)
sector. Deployment of this technology holds the potential
for large benefits—for example, connecting people and
firms to formal financial markets and linking farmers to
output markets. The African private sector has responded to
competition in this market in unprecedented ways. What is
behind the vibrancy of this market and what is the scope for
scaling up innovative uses of ICT in African countries?
Improving human development outcomes with inno-
vative Policies. Sub-Saharan Africa lags on all the health
and education MDGs. Yet, some countries have made sig-
nificant progress in improving health and education out-
comes. How have these countries achieved rapid progress?
What policies and approaches have worked and why? What
has been the role of public and private donors in improving
service delivery?
Each case study includes: a description of the achievement
and the elements that qualify the outcome as successful; an
assessment of the main policies, interventions, actions, and
other factors that contributed to the positive outcome; a pres-
entation of the lessons learned and the contribution to the
OVERVIEW 7
 discourse on African development; and, where, possible
insights on the transferability of the achievement and poten-
tial for scaling up the interventions and actions. Individual
case studies also examine the role of the key stakeholders—
the government, donors, or private investors—in facilitating
and promoting the achievement.
In examining government contributions to successful
outcomes, the focus is on the quality of economic policies
and actions—that is, whether public action relieved con-
straints to growth—whether they provided macroeconomic
stability, improved the investment climate, enhanced com-
petitiveness or eased access to markets, promoted human
development, or leveraged the global economy. How well
strategies and policies adjusted in response to changes in the
local and external economic environment is also addressed
in many of the case studies. The narrative avoids being pre-
scriptive because there are several policy combinations for
achieving desirable goals.
MAIN FINDINGS OF THE CASE STUDIES
There has been no shortage of narratives about Africa’s many
failures, giving rise to academic titles such as Africa’s Growth
Tragedy, or The Bottom Billion, not to mention relentless
media coverage of poverty, conflict, disease, and famine on the
continent. While the proximate causes of African countries’
failure to thrive are quite different—hyperinflation in
Zimbabwe, civil war in Sierra Leone, desertification in the
Sahel, and HIV/AIDS in South Africa, to name a few—they
can be boiled down to two main sources: market failures, such
as the common-property externalities associated with deserti-
fication, and government failures, often created by state inter-
vention to correct market failures. For instance, governments
throughout the world have central banks to manage monetary
policy because markets alone cannot do that; but Zimbabwe’s
central bank started issuing so much money to finance the fis-
cal deficit that the currency became worthless, resulting in
hyperinflation and many years of hardship.
Market failures can be corrected by implementing a tax
or subsidy or by generating a public good, such as improved
infrastructure or a better public health care system, for
which the government can take credit. Overcoming govern-
ment failures is more difficult, because the source of the
problem is usually powerful people who are benefiting from
the intervention. Correcting the problem involves under-
mining these people’s rents. In some cases, governments fail
to correct a market failure—another form of government
failure—because vested interests are benefiting from the
distortion (think of politically connected monopolists).
8 OVERVIEW
The 26 case studies in this book—all policy success sto-
ries from Sub-Saharan African countries—illustrate the
many ways Africa has avoided or overcome market and gov-
ernment failures. That these countries have surmounted
failures and generated economic and social progress bene-
fiting millions of poor people is testament to the innova-
tion, dynamism, and spirit of the African people. But the
case studies do more than inspire: they teach us about the
nature of government and market failures, and how they
can be prevented or corrected.
To learn from the studies, it is useful to classify them into
four categories: overcoming or avoiding massive government
failure, rebuilding or creating a government, rationalizing
government involvement in markets, and listening to the
people (table 1). Regardless of which category each of the
case studies falls into, as a group, they demonstrate that
the African landscape is dotted with success stories. The
sectoral and geographical diversity of the case studies
illustrates that there are many ways to overcome Africa’s
challenges. To paraphrase Tolstoy, African failures are all
alike; but each success is successful in its own way.
Overcoming massive government
failure or bad policies
In all the case studies in this group, existing policies were
standing in the way of growth, either at the sectoral or
economywide level. In Ghana’s cocoa sector, Rwanda’s cof-
fee sector, Kenya’s agricultural sector, Burkina Faso’s cotton
sector, and Tanzania’s wider economy, the willingness of
the government to allow liberalization has paid off. In
Uganda and Mozambique, postconflict reforms have paved
the way for impressive improvements in economic perfor-
mance. Across Sub-Sahara Africa, governments are grap-
pling with opening up the power sector to private sector
participation. Finally, the case study on Botswana proves
that the resource curse can successfully be avoided even in
an economy that is highly dependent on diamond mining.
All of these cases show that the timing and nature of
reforms is key to their success. But the specifics of each case
offer important lessons.
Long an integral part of Ghana’s economy, the cocoa
industry faced near collapse in the early 1980s, battered by
several forces. For one, Ghana was experiencing hyperinfla-
tion and an overvalued exchange rate (the market exchange
rate was approximately 44 times the official rate in 1983),
making selling its cocoa at official rates almost worthless. As
much as 20 percent of Ghana’s cocoa harvest was smuggled
into Côte d’Ivoire for export. At the same time, an aging tree
 Table 1 Categorizing Successes: Overcoming Government and Market Failure
Overcoming or avoiding massive
government failure
Rebuilding a government or creating
a government where none existed
Rationalizing government
involvement in markets
Listening to the people
Liberalization of the exchange rate and other reforms to revive the cocoa sector in Ghana
Removing barriers to trade and creating incentives for entrepreneurship in the coffee sector
in Rwanda
Liberalization of the fertilizer market in Kenya
Liberalization of the cotton sector in Burkina Faso
Facilitating private partnerships in the power sector across Sub-Saharan Africa through
independent power producers
Wide-ranging economic liberalization in Tanzania
Reforming the economy in a postconflict environment in Uganda and Mozambique
Good timing and good luck for diamond mining in Botswana
Rebuilding government following civil wars in Liberia and Sierra Leone
Using traditional systems for collective action in Somaliland
Development of a system of air, rail, and road transport and cold storage to support mango
exports in Mali
Provision of textile and apparel industry infrastructure in Lesotho
Catalytic government role in private sector development in Mauritius
Using input subsidies to improve agricultural output in Malawi
Provision of gorilla reserves to boost tourism revenues in Rwanda
Shifting the government role in the ICT sector from monopoly provider to regulator across
Sub-Saharan Africa
Success in malaria control across Sub-Saharan Africa
Performance-based financing in the health sector in Rwanda
Abolishment of fees to achieve free universal primary education in Uganda
Training and deploying extension workers to improve access to health care in Ethiopia
Lowering fertility through family planning programs
Developing new varieties of rice, NERICA, to increase yields and decrease food insecurity
Using Moneymaker pumps to support innovation and diffusion of technology in the
agricultural sector
Using mobile phones to improve financial access in Kenya via M-PESA

stock and the spread of plant diseases made Ghana’s cocoa
crop increasingly unproductive and investment in the cocoa
sector unattractive. Ghana’s proportion of global cocoa pro-
duction fell by half between the mid-1960s and the early
1980s, from 36 percent to 17 percent. Starting in 1983,
within the context of a wider economic recovery agenda
that devalued the Ghanaian cedi and increased the farm gate
prices paid to farmers for commodities (thus decreasing the
incentive to smuggle cocoa out of the country), the govern-
ment undertook a specialized program intended to revive
the cocoa sector. Among other reforms, the program com-
pensated farmers for replacing ailing cocoa trees (often with
better-performing hybrid varieties), shifted responsibility
for cocoa procurement to privately licensed companies, and
provided subsidies to increase the usage of fertilizer on
cocoa crops. Cocoa production and exports boomed. The
amount of cocoa produced per hectare doubled between
1983 and 2006 as a result of increased fertilizer application
and better production practices. A greater share of cocoa
sales is now passed on to the 700,000-plus cocoa farmers,
reducing poverty and improving prospects for cocoa-
producing households.
In Rwanda, where the vast majority of people depend on
agriculture for their livelihoods, major transformation of
the coffee sector came about more quickly than it did in
Ghana’s cocoa sector. As of the late 1990s farmers were pro-
ducing a small amount of mediocre-quality coffee that
attracted little attention from discriminating importers and
consumers, while the economic destruction of the country
during the genocide in 1994 remained a serious impediment
to markets. The first wave of reforms was introduced in
the late 1990s—several barriers to trade were removed, a
network of coffee-washing stations that allowed producers
to shift from semiwashed to fully washed (higher-value)
coffee was implemented, and incentives to foster entrepre-
neurship in the coffee industry were introduced. As a result,
both the quality and quantity of coffee produced in
Rwanda increased, allowing the government to launch,
in subsequent years, a successful strategy to boost Rwandan
producers’ participation in the international specialty-
coffee market. Between 2003 and 2008, the average export
price of green coffee originating in Rwanda nearly doubled,
from $1.60 to $3.10 a kilogram, leading to important
improvements at the household level. The coffee washing

OVERVIEW 9
 stations, which had created 4,000 jobs as of 2006, also
appear to have led to improved economic cooperation
among people using them.
Like cocoa in Ghana, cotton has long been integral to the
economy of Burkina Faso. Cotton accounted for 60 percent
of the country’s exports between 1994 and 2004 and
contributed to Burkina Faso’s good economic performance.
A series of institutional reforms implemented in the cotton
sector during the 1990s and 2000s—including the formation
of a national cotton union and the partial privatization of
the national cotton parastatal, SOFITEX—have succeeded in
opening the sector. The Burkinabe reform model is unique
among Sub-Saharan African countries, because it addressed
government failures and local realities within the existing
institutional framework, adopted reforms using a cautious,
piecemeal approach, and built the capacity of producers and
upgraded institutions within the commodity chain while the
government withdrew from most of its activities. But it is
also clear that Burkina Faso’s dependence on cotton makes
economic growth and performance in the country quite
volatile, because cotton earnings are dependent on interna-
tional markets. In the years ahead, the country will need to
seek avenues by which to diversify the economy.
In Kenya, as in Sub-Saharan Africa as a whole, fertilizer
use traditionally has been quite low, meaning that overall
productivity in the agricultural sector is also quite low.
Combined with the price volatility brought about by
unpredictable weather and poor infrastructure, this low
productivity has contributed to food insecurity in Kenya.
For decades, the Kenyan government addressed this food
insecurity through direct, monopolistic involvement in the
markets for a wide array of agricultural products. This
changed in the 1980s, when the government began relaxing
its hold on agricultural markets, allowing the private sector
to compete with state agencies and easing trade restrictions
on fertilizer and maize. A more complete liberalization of
the fertilizer market occurred in 1993. This change in policy,
coupled with liberalization of the foreign exchange regime
in 1992, encouraged the entry of a significant number of
private sector firms in importing, wholesaling, distribution,
and retailing of fertilizer. Government price controls and
import licensing quotas were ultimately eliminated, and fer-
tilizer donations by donor agencies were phased out. Kenya
now stands as a notable departure from the Sub-Saharan
African average in terms of fertilizer usage, which almost
doubled between 1992 and 2007. Much of the increased use
has been among smallholder farmers. Largely as a result of
the increase in fertilizer usage, maize yields in Kenya
increased 18 percent over 1997–2007.
10 OVERVIEW
Efforts to overcome bad policies in the power sector have
been effective in some Sub-Saharan African countries. The
case study on independent power producers (IPPs) focuses
on the seven countries that have had the most experience
with IPPs. As of the early 1990s, virtually all major power
producers in Sub-Saharan Africa were publicly owned. They
had been performing poorly for decades, and public financ-
ing for new projects (from both domestic and international
sources) was drying up. Governments needed to develop a
new tactic. Starting in Côte d’Ivoire in 1994, they experi-
mented with opening power generation to the private sec-
tor. Ghana, Kenya, Nigeria, Senegal, Tanzania, and Uganda,
among other countries, followed suit shortly thereafter.
While the presence of IPPs has not solved the electricity
deficit on the continent—only 25 percent of the population
has access to electricity—IPPs have added several gigawatts
of capacity to Sub-Saharan Africa’s electric grid, comple-
menting incumbent state-owned facilities. It is also clear
that a sound policy framework (namely, adequate legisla-
tion), a good investment climate, and local availability of
cost-competitive fuel can all contribute to the success of
IPPs. When these elements are not in place, IPPs are much
more likely to be untenable.
The case study on Tanzania addresses long-term market
reforms that have gone far beyond a single sector. Following
independence, the country’s economy languished for more
than two decades, held back by loss-making public enter-
prises, deteriorating infrastructure, and mismanagement
of terms of trade and weather shocks, among other
things, while poverty increased. Starting in 1986, a series of
structural reforms focusing on liberalization was imple-
mented, many of which were similar to the reforms put in
place by Uganda in the same time period (see below). Trade
and exchange rate regimes were liberalized; marketing
boards for agricultural products were dismantled; and
parastatals, the financial sector, and the civil service began
to be reformed. The reform cycle continued into the 1990s,
supported by a significant amount of debt relief from inter-
national donors, which freed budgetary resources for other
purposes. By the mid-1990s Tanzania was on the cusp of a
decade-long growth take-off, with real GDP growth averag-
ing around 6 percent from 1996 to 2008. Despite a difficult
external environment, Tanzania has maintained economic
stability in recent years. Aside from the fact that fiscal and
monetary authorities have succeeded in putting the econ-
omy on much better footing than existed decades ago, two
other factors—low reliance on exports for growth and low
levels of foreign exposure within the banking sector—
allowed Tanzania to weather the global financial crisis
relatively unscathed. And while poverty among Tanzania’s
population remains high, the country is on track to meet
several of the targets laid out under the Millennium
Development Goals.
In Uganda sustained, carefully sequenced macroeco-
nomic reforms in a postconflict environment have produced
remarkable outcomes. At the end of decades of political
instability and civil war in 1986, Uganda’s economy was in
disarray and its population was struggling with a very high
incidence of poverty. The reform agenda, begun in 1987 and
implemented in stages, focused on stabilization, liberaliza-
tion, and structural reforms: prices controls were loosened,
a floating exchange rate regime was adopted, marketing
boards were abolished, parastatals were abolished or priva-
tized, the civil service was reformed, and efforts were made
to stimulate private investment. These steps were followed
by a wide-ranging poverty reduction plan that started in
1997. The performance of Uganda’s economy in the 1990s
and 2000s suggests that the reform policies worked. Growth
averaged 7.7 percent a year over 1997–2007. Increased inter-
national confidence in the economy spurred substantial aid
and foreign direct investment inflows and reversed capital
flight. Poverty figures have also improved dramatically, so
much so that Uganda will be one of the few Sub-Saharan
African countries to achieve the first Millennium Develop-
ment Goal of halving extreme poverty by 2015. At the same
time, the Uganda case strikes a cautionary note: despite
strong growth, income inequality has increased and the for-
mal sector employment rates are very low, particularly
among youth.
Having emerged from 16 years of devastating civil war in
1992 as one of the poorest countries in the world and with
the second-lowest Human Development Index score, the
economic situation in Mozambique was bleak. As in
Uganda, Mozambique implemented a series of macroeco-
nomic and structural policy reforms shortly after the war
ended, in addition to encouraging support from the interna-
tional donor community. The turnaround in Mozambique
between 1993 and 2009 was profound: growth averaged
more than 8 percent a year (compared to 0 percent from
1981 to 1992), the poverty rate dropped dramatically, and
social indicators improved. Double-digit growth has been
observed in mining, manufacturing, construction, electric-
ity, gas, and water. Sound macroeconomic management
allowed Mozambique to attract an increasing amount of
foreign direct investment, which increased from an average
of 1.5 percent of GDP in 1993–98 to 5.2 percent of GDP in
1999–2010. While progress has been impressive, in the years
ahead, Mozambique will need to determine how to create
more jobs (particularly outside the agricultural sector,
where most people are still employed) and widen the pro-
ductive base in order to create an economy in which wealth
is more broadly shared.
Policy choices in Botswana, like those in Tanzania, have
been key to long-term economic progress. At the time of
independence in 1966, Botswana’s per capita income was
among the lowest in the world. Many of its human devel-
opment indicators were equally abysmal: life expectancy
was 37 years, and the primary school completion rate was
less than 2 percent. The country’s outlook was not encour-
aging. But Botswana’s economic performance in the years
since has proven those expectations wrong: per capita
income growth, for example, averaged 7 percent a year
between 1966 and 1999. Without a doubt, Botswana met
with a good bit of luck along the way—the discovery of
large deposits of diamonds just after independence. In the
decades since, diamonds have fueled solid economic
growth and a rapid increase in per capita annual income.
But effective, transparent governance and good economic
management have been crucial to Botswana’s success; they
helped Botswana avoid the resource curse experienced by
other Sub-Saharan African countries. In addition, the poli-
cies Botswana did not adopt also appear to have con-
tributed to its economic success. Botswana did not follow
an import substitution policy, and it did not expand state-
owned enterprises, which employ only a small percentage
of the workforce. In the ensuing decades, economic and
social indicators in Botswana outperformed those of Sub-
Saharan Africa as a whole by a wide margin. Life
expectancy in Botswana was 60 years as of 1990, 10 years
above the African average, while the under-five mortality
rate had fallen to about 45 per 1,000, compared with 180
for Africa as a whole.4
Creating or repairing government
The second set of case studies details how countries have
recovered from what is arguably the biggest government
failure of all—civil conflict. In Liberia and Sierra Leone,
governments have been successfully rebuilt from the rubble
of devastating wars. In Somalia, on the other hand, which
has existed without an internationally recognized central
government since 1991, traditional power structures have
been used in the breakaway region of Somaliland to
strengthen infrastructure.
A long, violent conflict that ended in 2003 left Liberia in
a state of collapse. Corruption was rampant, external debt
was approximately 800 percent of GDP, the government’s
OVERVIEW 11
 ability to manage public finances had essentially disinte-
grated, and the Liberian people received among the lowest
amounts of per capita public spending in the world—
approximately $25 a year. In the years since, the country’s
reform agenda has vastly improved the governance and eco-
nomic environment. Liberia’s situation took a solid turn for
the better at the beginning of 2006. Under the direction of
the newly elected president, Ellen Johnson Sirleaf, and in
conjunction with a host of international donors, Liberia
embarked upon a broad-ranging recovery and reform plan.
Improving revenue collection and establishing an effective
national expenditure process were the initial areas of focus.
A unique feature of the program was its cosignatory
arrangement, under which no major public financial trans-
actions could take place without first being scrutinized by
both a Liberian manager and an international advisor. This
element of balance was fundamental to the functioning of
the overall reform program. While the outcomes have been
generally positive—revenue collection and expenditure
management has improved dramatically—a final lesson is
that a lot of capacity building still has to be done in the area
of economic governance in Liberia.
Neighboring Sierra Leone was in a situation similar to
that in Liberia. More than a decade of civil war in Sierra
Leone came to a close in 2002, leaving the little infrastruc-
ture the country had in shambles, measures of human
development among the worst in the world, and capacity for
economic and political governance virtually nonexistent.
The process of restoring (and indeed, improving) the system
of political governance in Sierra Leone was a crucial part of
the country’s postconflict reform program, because margin-
alization of people living outside the capital, Freetown, was a
main cause of the conflict. A national reform program,
begun in 2004, focused on devolving political, fiscal, and
administrative power to local governments. In some respects,
the reform process has been quite successful—access to
quality health services has improved dramatically, for one,
and previously marginalized groups, such as women and
ethnic minorities, have benefited from the new space for
political participation. But in other areas positive results are
less clear. In addition, recent developments within the cen-
tral government indicate there may be increasing pressure
to pull back from the decentralization, potentially reducing
the future successes of the initiative.
Finally, the case study on Somaliland is an illustration of
traditional structures playing the role of government where
no formal system of government existed. Following the col-
lapse of the central government of Somalia in 1991, much of
the country fell into the domain of bandits and warlords,
12 OVERVIEW
making transporting goods through and out of the country
a risky ordeal. But Somaliland, which unilaterally declared
its secession from Somalia the same year the central govern-
ment collapsed, has made admirable progress in improving
its infrastructure to foster trade—namely, in expanding the
port of Berbera, on the Gulf of Aden, and in ensuring the
usability and security of the highway that leads to the port.
Travel on the road is now reliable enough that Ethiopia
transports a large portion of its exports along it before ship-
ping them from Berbera. Somaliland’s success in building
up its infrastructure stands in sharp contrast to the situation
in the rest of Somalia, which remains tense and dangerous.
Factors contributing to Somaliland’s success were twofold:
use of the traditional social structures to control violence,
and successful political cooperation in order to instill the
level of security needed to allow development and operation
of the port. While Somaliland is an extreme case, it is
undoubtedly representative of scores of cases across Sub-
Saharan Africa in which local communities succeed in pro-
viding leadership the official government is unable or
unwilling to provide.
Rationalizing government involvement in markets
The third set of case studies illustrates how governments
have successfully intervened to correct market failures—and
no more. In the cases of mangoes in Mali, gorilla tourism in
Rwanda, and apparel production in Lesotho, the respective
governments stepped in to provide the elements necessary
for the private sector to thrive. Responding to the possibil-
ity of profits, private entrepreneurs took on the bulk of the
work to develop the sectors in question. In Malawi meas-
ured government involvement in agriculture has produced
positive outcomes, while in Mauritius strategic government
involvement in several sectors over the course of three
decades played a catalytic role in the country’s private sector
development. Across Sub-Saharan Africa, the government’s
pullback from the power and information and communica-
tion technology sectors has allowed more people to take
advantage of those services. The ICT sector, especially, has
experienced a dramatic take-off. In all of these cases, the
governments provided services essential for private sector
success and then stepped back to allow private sector actors
to carry out sectoral development. Even in difficult sectors
such as health, initiatives that balance private and public
sector involvement have produced truly impressive results,
as discussed in the case study on malaria control.
In the 1970s Mali, where the vast majority of the popula-
tion works in agriculture, began to explore opportunities to
export its fresh mangoes. Malian mangoes quickly found a
market in specialty stores in France and elsewhere in
Europe. But Mali is landlocked, so traditionally, it either
depended on neighboring countries for road, rail, and port
infrastructure for export purposes or shipped goods from
within its own borders via expensive air freight. By the early
1990s, with demand for mangoes in Europe increasing and
countries such as Brazil (where shipping costs were lower
because of good ocean transport options) becoming bigger
producers, Mali grew increasingly uncompetitive in the
global market for mangoes. Less expensive, more efficient
methods of transportation were needed. Starting in 1995
the government in partnership with the private sector—
farmers, farmer groups, and other businesses—launched an
innovative response: a multimodal supply chain overhaul
that used temperature-controlled containers to transport
mangoes and improved road, rail, and sea links with neigh-
boring countries. The level of government involvement in
the mango sector turned out to be ideal, and the efficiency
improvements have led to a host of positive outcomes: a 150
percent increase in the price mango producers receive for
their products, a 1,000 percent increase in the tonnage of
mangoes exported between 1993 and 2008, and a reduction
in the average transit time for mangoes between Mali and
Europe, from 25 to 12 days.
Another landlocked country in Sub-Saharan Africa,
Lesotho, was also struggling to export in the 1990s, thanks
to its limited transportation infrastructure, underdeveloped
factor markets, and inadequate backward and forward
industrial linkages and technical expertise. The solution
came in the form of an international mandate, the U.S.
African Growth and Opportunity Act (AGOA). In the early
2000s Lesotho launched a series of aggressive investment
and export promotion strategies that positioned it to capi-
talize on the apparel industry benefits of AGOA. One
notable aspect of Lesotho’s experience is the government’s
decision to combine an apparel industry competitiveness
initiative with a series of early-stage incentives for investors
(they were offered publicly owned factory shells at subsi-
dized rents). The tactic has yielded positive results—namely,
that Lesotho’s apparel exports to the United States, at $177
per capita in 2009, are higher than in any other apparel-
producing country in Sub-Saharan Africa.
In Malawi a government initiative in the agricultural sec-
tor has had positive results within just a few years. Launched
by the Malawian government in 2005 in response to severe
food security difficulties in the early 2000s, the agricultural
input subsidy program was intended to reverse the low pro-
ductivity and high price of maize, a staple food for the
country and an important source of jobs. While interven-
tion in the maize market, including input subsidies, has
been a longstanding—though often contentious—feature
of government and international donor strategies to pro-
mote agricultural productivity and food security in Malawi,
the 2005 program went further than past efforts in its scale
and scope. Vouchers for maize fertilizers were provided to
more than half of Malawi’s farm households. Additional
vouchers were distributed for improved maize seeds and
tobacco fertilizers. The core objective of the program was
twofold: improving farmers’ ability to achieve food self-
sufficiency, and boosting farmers’ incomes through
increased crop production. Without a doubt, Malawi’s input
subsidy program has had positive impacts: maize produc-
tion is estimated to have increased by 26–60 percent under
the program, while food availability has improved and real
wages in the agricultural sector have increased. But doubts
about the effectiveness and appropriateness of the program
have also been raised—namely, that the fiscal costs of the
program are high, as are its opportunity costs (in terms of
crowding out other needed public investments), and that it
may be impeding the development of sustainable commer-
cial agricultural input markets.
With the specter of the genocide still looming in the back-
ground in 1994, the Rwandan government embarked upon a
national recovery strategy that included the redevelopment
of the tourism sector. The goal was to focus on attracting
high-end, conservation-oriented tourists interested in view-
ing Rwanda’s gorilla population. The strategy also included
an international marketing campaign intended to improve
Rwanda’s image in the world. Importantly, it also called for
near-complete privatization of the hotel and leisure sector.
The results have been impressive. Tourism has emerged as
Rwanda’s top foreign currency earner, ahead of the coffee
and tea sectors. The number of visitors to Rwanda’s national
parks has increased exponentially—from 417 in 1999 to
43,000 in 2008. Moreover, it is clear in the case of Rwanda
that initial actions by the government to revive the tourism
sector have encouraged active involvement of the private
sector: 86 percent of all new tourism-related projects in the
country are locally owned.
The case study on Mauritius discusses government
involvement in markets not over several years but several
decades. Much has been made of the Mauritian “miracle,”
whereby the island transformed itself from a poor, sugar-
dependent economy at independence in 1968 into what it is
today: a fast-growing, much more diversified economy with
one of the highest per capita income levels in Africa. Along
the way, human development indicators such as life
OVERVIEW 13
 expectancy (73 years in 2008, compared to 62 years in 1970)
have registered impressive improvements. And unlike in
many other fast-growing economies, income inequality in
Mauritius has declined solidly over the past 40 years. Aside
from prudent fiscal, monetary, and exchange rate policies,
Mauritius’ economy has benefited from vibrant partnership
between the public and private sectors. Among the most vis-
ible of the results of these partnership efforts has been the
establishment of export processing zones (EPZs) to push
along development of the light manufacturing sector in the
1980s and policies supporting growth in the financial ser-
vices and ICT sectors in recent decades. Another important
key to success in Mauritius has been the country’s ongoing
efforts to forge consensus between the Franco-Mauritian
business elite and the Indo-Mauritian political elite.
In the health sector, too, an appropriate level of govern-
ment participation in markets has produced positive out-
comes in Sub-Saharan Africa. One example is in the battle
against malaria. Governments, communities, donors, and
individuals are increasingly coordinating their responses to
the disease, reducing duplicative efforts, increasing the
capacity to mobilize resources, raising awareness of the
problem, and creating a network of technical and imple-
mentation experts. Efforts under the umbrella of the Roll
Back Malaria Partnership, for example, have been a key in
reducing the incidence of malaria in numerous Sub-Saharan
African countries, such as Eritrea, Ethiopia, Rwanda, and
Zambia by promoting the use of insecticide-treated bed
nets, managing malaria vector breeding sites, and provid-
ing widespread diagnostic and treatment services. In
Eritrea, for example, malaria morbidity dropped from
about 100,000 cases in 2000 to about 8,000 in 2008. In
Zambia the percentage of under-five children with
malaria parasite prevalence fell from 22 percent in 2006 to
10.2 percent in 2008; the number of deaths caused by
malaria declined 47 percent over the same years. These
emerging successes notwithstanding, more remains to be
done to control malaria in Sub-Saharan Africa.
Finally, the information and communication technology
revolution that has occurred across Sub-Saharan Africa
since the late 1990s is the result of government moving out
of the way of the private sector. With demand for mobile
phones increasing, most governments in Africa switched
from being the monopoly provider (as they were for land-
lines) to being only the regulator in the mobile phone
industry. As a result, the private sector was able to move in,
increasing competition and reducing costs. The number of
mobile phone subscribers grew exponentially, from 4 mil-
lion in 1998 to 259 million in 2008, dramatically improving
14 OVERVIEW
the ability of people across Sub-Saharan Africa to commu-
nicate. Phone services are now affordable to the majority of
Africans rather than a privileged few. Private sector involve-
ment in the ICT sector has also allowed innovative, mobile-
based services in the banking, agricultural, and health sectors
to take hold quickly.
Listening to the people
The final category of case studies examines policies and
innovations that were successful in part because end users
were consulted in the process of developing them. Two of
the case studies in this section discuss how the provision of
tangible products—human-powered water pumps and
new varieties of high-performing hybrid rice—have con-
tributed to a reduction in food insecurity across Sub-Saharan
Africa. Four other cases studies detail important successes
at the country level: abolishing school fees to achieve uni-
versal primary education in Uganda; providing health ser-
vices in underserved rural areas in Ethiopia; implementing
a performance-based financing system to improve health
coverage in Rwanda; and reducing the cost of transferring
money in Kenya through M-PESA, a mobile-phone-based
electronic payments system. In some cases, similar innova-
tions had been previously attempted but failed. The differ-
ence this time around was the way governments went
about policy making. Rather than assuming they under-
stood what worked well for their people and implementing
policy from the top down, governments elicited feedback
from the public in the course of formulating and imple-
menting the policy. They adopted bottom-up approaches.
As a result of being included in the policy-making process,
people were much more likely to benefit from the final
product or service. Finally, the case study on family plan-
ning programs illustrates that the responsiveness to users
of family planning services is a prerequisite to reducing
fertility.
Founded in Kenya in 1991, KickStart International’s mis-
sion is to promote economic growth and employment cre-
ation in Sub-Saharan Africa by developing and promoting
technologies that can be used to establish and operate prof-
itable small-scale businesses, including its range of low-cost,
human-powered pumps that allow farmers to irrigate their
land. In turn, pump users are able to increase their yields
(providing more food for their families), earn additional
income, and create agricultural jobs. An important part of
KickStart’s model is the participatory fashion in which
products are developed, with farmers advising on marketing
and, in the case of the SuperMoneyMaker pump, the design
itself. This element has been key in attracting pump cus-
tomers and, as a result, improving agricultural returns.
That participatory spirit also contributed to the success-
ful uptake of the New Rice for Africa (NERICA) hybrid
varieties in the 1990s. The rationale behind developing the
hybrids was quite clear. Though rice is the third-highest
source of caloric intake and the fastest-growing food staple
in Africa, up to 40 percent of rice is imported. From the
start, the developers of NERICA, the Africa Rice Center and
a consortium of partners, intended for it to increase food
security and farmers’ incomes. Importantly, the team con-
sulted extensively with farmers in the course of developing
the new varieties. That extra step ensured active adoption of
the new varieties in a relatively short time frame. Initial
empirical evidence from five Sub-Saharan African countries
indicates that NERICA has had positive impacts on yields—
especially in Benin and The Gambia—and on household
income. Female farmers, in particular, have profited from
cultivating NERICA varieties.
In Uganda free primary education was abandoned in the
1980s because the government’s money was not reaching
schoolchildren. Budgetary leakages were ubiquitous, and
teacher performance abysmal. But after more than a decade,
it was clear that charging school fees was not working either.
When Uganda reintroduced free primary education in the
1997, it not only provided an incentive for parents to send
their children to school but also publicized the amount of
money each school district was receiving as a way for the cit-
izens to hold the government accountable for the program.
By quantitative measures, the universal primary education
initiative in Uganda has been a resounding success. Net pri-
mary enrollment rose from 57 percent to 85 percent in a
single year, from 1996 and 1997, and to 90 percent in 1999.
Importantly, poor children, girls, and rural residents have
benefited disproportionately from the increase in access.
Over time, the quality of primary education in Uganda has
also improved: between 2003 and 2010, for example, the
numeracy rate among students at the Primary 3 level rose
from 42 to 72 percent, while the literacy rate increased from
36 to 58 percent. That said, there is still considerable space
for improvement in Uganda’s education sector, particularly
in reducing overcrowding in schools, ensuring adequate
supply of teachers and materials, and reducing the number
of over-age students in primary schools.
Cases in which governments acted with sensitivity to the
people have also brought dramatic improvements in the
health sector. In Ethiopia, a country of 80 million people,
mostly in rural areas, more than half lived more than 10
kilometers from a health services facility as of 2005. Yet the
rural poor in Ethiopia, where maternal and child mortality
ratios were among the highest in the world and the number
of trained health workers was inadequate, desperately
needed better health services. Seeking innovative ways to
respond to this challenge, policy makers came upon the idea
of creating a national network of health extension workers
who could address the basic health needs of the rural popu-
lation. The program, launched in 2003, involved training
and deploying more than 34,000 (predominantly female)
community health workers who provide essential services
covering hygiene, disease control and prevention, family
planning, and health education. Although full implementa-
tion of the program was completed only in 2010, the out-
comes thus far have been extremely promising. Childhood
vaccination coverage has increased dramatically: 62 percent
of children were fully immunized as of 2010, representing
an average annual increase of 15 percent since 2006. Prena-
tal and postnatal maternal health coverage has improved.
Women who previously desired family planning services
but were unable to find them now have access to those ser-
vices. Use of insecticide-treated bed nets increased 15-fold
between 2004/05 and 2009/10, so that 68 percent of house-
holds in malaria-affected areas are now protected by at least
one bed net or indoor residual spray; the reduction in
malaria incidence in recent years is attributable to expanded
coverage of these malaria interventions.
In Rwanda, too, listening to what people need has
worked in the health sector. In addition to broad health sec-
tor reform and fiscal decentralization, Rwanda made use of
two innovative policy tools—community-based health
insurance and performance-based financing—to increase
access to health services. The first policy made health ser-
vices more available in rural areas and available at a lower
cost. Under the second policy, Rwanda pays doctors a bonus
depending on the number of children immunized or the
number of pregnant mothers examined, more closely link-
ing rewards with performance. Rwanda’s innovative policy
tools have contributed to the profound improvements in
health indicators in recent years. Under-five mortality, for
example, dropped from 196 to 103 per 1,000 live births
between 2000 and 2007, and the maternal mortality ratio
fell by more than 12 percent each year from 2000 to 2008.
Rwanda is on track to reach the maternal health Millen-
nium Development Goal by 2015. Improvements among
the most vulnerable segments of the population have been
particularly evident.
In crafting family planning programs, being responsive
to the needs of users is a prerequisite to reducing fertility.
Sub-Saharan Africa has the highest total fertility rate in the
OVERVIEW 15
 world, but some countries in the region are undergoing
dynamic and unprecedented fertility transitions. In coun-
tries in which the greatest progress has been made, some of
the key ingredients of success have been a high level of
political commitment, strong country-level institutions,
and effective family planning service delivery strategies.
Family planning programs that have delivery points
throughout the country; provide a range of contraceptive
methods; ensure easy availability of contraceptives; adopt a
reproductive health approach; and reach adolescents, men,
and unmarried people are most likely to accelerate progress
toward fertility decline in Africa. Many countries in Africa
have tried community-based distribution of contraceptives
to extend family planning to hard-to-reach populations,
particularly in rural areas. Community depots, mobile clin-
ics, women’s groups, and both paid and volunteer village
health workers are some modes of service delivery utilized
by such programs.
The phenomenal success of M-PESA’s mobile payments
system is the quintessential example of (the private sector, in
this case) providing a service that poor people in Kenya
sorely needed—the ability to transfer small sums of money
to people in remote areas at a low cost. Built on a mobile
phone platform, M-PESA filled that niche, enabling cus-
tomers to send money and store money through a simple
interface. The service has been wildly successful. Launched
in 2007, the number of subscribers surpassed 9 million in
late 2009. Recent figures indicate that M-PESA handles $320
million in person-to-person transfers a month, or roughly
10 percent of Kenya’s GDP on an annualized basis.
WHY DID REFORM HAPPEN AND WHY WAS
IT EFFECTIVE?
The above findings beg two important questions: what
prompts change in policies, and why did policy reforms
work? In all but one of the cases of correcting massive gov-
ernment failure, existing policies were standing in the way of
growth. How then did change come about? How were the
obstacles to correcting government failure overcome? To
address these issues one needs to understand why failed poli-
cies were present to begin with. Central to this issue is polit-
ical context—that is, the political economy factors that
induced the prior bad policies. There is a vast literature
examining the political economy of poor policies, and study-
ing it could help countries recognize when reforms could
work and how to design reforms. Although an assessment of
the political economy dimension is outside the scope of this
study, some observations can be made based on the findings.
16 OVERVIEW
From a political economy perspective, poor policies per-
sist because the preferences of those in power are not
aligned with those of the public, and the former use bad or
inefficient policies for political purposes or financial gain,
or both (either politicians are beholden to interest groups or
politicians create interest groups that are beholden to
them). Individuals and groups who expect to lose economic
rent from policy correction are likely to resist change or
weaken its effects. Accountability mechanisms—checks and
balances—on policy makers are designed to reduce incen-
tives for opportunistic behavior and distortionary policies.
But coordination and collective action problems prevent the
establishment of strong accountability mechanisms. The
weaker the constraints on politicians, the less likely reforms
are to occur. Conversely, where political constraints are
strong, bad policies will not persist, so the scope for reform
to have a major improvement is low (Acemoglu et al. 2008).
However, policy change can occur in an environment of
weak political constraints when there is a big change in
political power.5 Such a change disrupts the existing status
quo and shifts power from existing interest groups to new
ones and possibly to a broader group of people. That is what
seems to have happened in the case of several countries is
the study.
For example, in Ghana there was a shift from military
rule to democracy. Under military rule the government
had an incentive to keep agricultural and import prices
low because its power base was the urban elite, including
the army. But when democracy came, politicians needed
the rural vote, so it was in their interest to liberalize the
exchange rate and allow agricultural prices to rise. In
Rwanda the situation was different. The liberalization of
the coffee sector was part of the postgenocide govern-
ment’s attempt at restoring growth and rebuilding trust in
government among all the people.
Likewise, in Uganda and Mozambique, both of which
had been devastated by decades of political instability and
civil war, major political reforms occurred in a postconflict
environment. In both cases the new stability allowed new
governments to adopt broad-based reforms to reverse eco-
nomic decline and boost economic development.
Botswana, on the other hand, represents a case where the
country was able to avoid bad policies after independence,
despite the discovery of resource riches. One reason why the
elite did not capture resource rents was the presence of tra-
ditional institutions—the Tswana tribal tradition of
consultation, known as kgotla—which emphasized that the
government exists to serve the people and promote develop-
ment and is not the instrument of one group or individuals
for the purpose of getting hold of the wealth. Tswana tradi-
tion also respected private property; the fact that many of
the tribal leaders who helped usher in modern government
were also large cattle owners may have reinforced this
respect. These traditional institutions promoted respect for
property rights and the rule of law, reducing incentives for
distortionary policies.
Although transformation or creation of functioning
institutions takes time, Radelet (2010) finds strong evidence
of improved governance and more accountable govern-
ments in African countries that have achieved steady eco-
nomic growth and lower poverty rates since the mid-1990s.
As noted above, stronger constraints on politicians reduce
incentives for inefficient policies.6 It would thus appear that
strengthening institutions can be extremely effective in
bringing about change. Many of these countries are also
identified in our study as experiencing successful growth
episodes (table 2).
The range of reforms—exchange rate liberalization,
opening up of trade, reduction of barriers to market entry,
and liberalization of input and output markets—are not
new. But why do these reforms work in some cases and fail
in many other instances? Country context is one reason: it is
central to understanding what works. A few insights are
provided by the case studies. One possible factor could be
that the reform was part of a broader set of reforms. This
meant that successful policy reform in one area was not
being largely offset or negated by reforms elsewhere—what
Acemoglu et al. (2008) call a “seesaw” effect. The seesaw
occurs when successful policy reform in one area is largely
offset by reforms elsewhere as powerful interest groups
opposed to the reforms attempt to thwart their effect.
Simultaneous reforms that were complementary meant that
the effect of any one reform was likely to be enhanced.
Another factor is the level of commitment to reform.
Here, it is important to consider what prevents the reform
from being reversed. The case studies on growth find that
policy makers were generally able to commit to appropriate
reforms—especially macroeconomic stabilization. One
could reasonably argue that a benign global environment of
strong growth, rising commodity prices, and low interest
rates provided an enabling environment for sensible macro-
economic and structural reforms.
CONCLUDING REMARKS
Whether viewed through aggregate indicators or the case
studies in this book, Africa’s growth and development over the
past decade has been impressive. The question on everyone’s
mind is: Will it be sustained? By showing that progress has
come about through a combination of policy reform and
active government interventions—balancing market failure
and government failure—the 26 case studies included here
give cause for optimism. If, as seems to be the case, African
governments are reforming large, distortionary policies, and
replacing them with selective interventions where there is gen-
uine market failure and, more important, intervening with the
feedback of the ultimate beneficiary to avoid government fail-
ure, then there is a good chance that the continent’s strong
economic performance of the last decade will be sustained.

NOTES
Table 2 Countries with Strong Economic
1. Several studies find a shift in trend beginning in the
Performance Saw an Improving Trend
mid-1990s.
in Institutions
2. World Bank 2005; Growth Commission 2008; IFPRI
Freedom House Polity IV Score
1.0 = best score 10 = best score
2009; Center for Global Development 2007.
7.0 = worst score –10 = worst score 3. The National Bureau for Economic Research Africa
Country 1989 2008 1989 2008 Project, which is ongoing, identifies and analyzes a large
Botswana 1.5 2.0 7 8 number of economic development successes in the region—
Burkina Faso 5.5 4.0 –7 0 35 cases have been already selected for in-depth study and
Liberia 5.5 3.5 –6 6 40 are anticipated—to better understand the factors behind
Mozambique 6.5 3.0 –7 6 the positive experiences and to evaluate the sustainability of
Rwanda 6.0 5.5 –7 –3
Sierra Leone 5.5 3.0 –7 7
the successes and their transferability to other African coun-
Tanzania 6.0 3.5 –6 –1 tries. The study covers four broad topics: macroeconomic
Uganda 5.0 4.5 –7 –1 dimensions of growth; microeconomic aspects of growth;
the intersection of health and growth; and cross-regional
Source: Radelet 2010.
Note: Polity IV measures qualities of executive recruitment con-
comparisons. The full findings of the 4.5 year project
straints on executive authority, and political competition. Freedom (2007–12), which is being supported by the Gates Founda-
House measures political rights and civil liberties. tion, will be available in a few years.

OVERVIEW 17
 4. Life expectancy and health indicators have deteriorated
in the face of the HIV/AIDS epidemic.
5. Politicians can also adopt reforms when their earlier
policies prove to be utterly ruinous to the economy and
changing course seems to be the only viable option to avoid
political change. Jones and Olken (2007) find empirical evi-
dence that assassination of autocrats affects institutional
change—specifically, a move toward democracy.
6. Where there is competition, inefficient institutions will
be eliminated ( Kingston and Gonzalo Caballero 2008).
REFERENCES
Acemoglu, Daren, Simon Johnson, Pablo Querubin, and
James A. Robinson. 2008. “When Does Policy Reform
Work? The Case of Central Bank Independence.”
Brookings Papers on Economic Activity (Spring):
351–418.
Arbache, J., D. Go, and J. Page. 2008. “Is Africa’s Economy at a
Turning point?” In Africa at a Turning Point? Growth, Aid,
and External Shock ed. D. Go and J. Page, 13–85. World
Bank. Washington, DC
Arbache, J. and J. Page 2007. “More Growth or Fewer
Collapses? A New Look at Long-Run Growth in Sub-
Saharan Africa.” Policy Research Working Paper 4384.
World Bank, Washington, DC.
Block, S. 2010. “The Decline and Rise of Agricultural Pro-
ductivity in Sub-Saharan Africa Since 1961.” NBER
Working Paper 16481, National Bureau for Economic
Research, Cambridge, MA.
Center of Global Development. 2007. Case Studies in Global
Health: Millions Saved.” Washington, DC.
Cohen, Jessica, and William Easterly, eds. 2009. What Works
In Development?: Thinking Big and Thinking Small.
Brookings Institution, Washington, DC.
Fuglie, Keith. 2010. “Agricultural Productivity in Sub-Saharan
Africa.” U.S. Department of Agriculture, Washington, DC.
18 OVERVIEW
Commission on Growth and Development. 2008.
The Growth Report: Strategies for Sustained Growth
and Inclusive Development. Washington, DC: World
Bank.
Devarajan, Shantayanan, David Dollar, and Torgny
Holmgren. 2001. Aid and Reform in Africa: Lessons from
Ten Case Studies. World Bank, Washington, DC.
IFPRI (International Food Policy Research Institute). 2009.
Millions Fed: Proven Successes in Agricultural Develop-
ment. Washington, DC.
Jones, B., and B. Olken. 2007. “Hit or Miss? The Effect of
Assassinations on Institutions and War.” NBER Working
Paper 13102. National Bureau for Economic Research,
Cambridge, MA.
Kingston, C., and G. Caballero. 2008. “Comparing Theories
of Institutional Change.” Journal of Institutional Econom-
ics 5: 151–80.
Radelet, Stephen. 2010. Emerging Africa: How 17 Countries
Are Leading the Way. Washington, DC: Center for Global
Development.
UNIDO (United Nations Industrial Development Organi-
zation). 2009. Industrial Development Report 2009:
Breaking In and Moving Up. Vienna: UNIDO.
World Bank. 2000. Can Africa Claim the 21st Century?
Washington, DC: World Bank.
———. 2005. Scaling Up Poverty Reduction: Learning
and Innovating for Development: Findings from the
Shanghai Global Learning Initiative. Washington, DC:
World Bank.
———. 2008. World Development Report 2008: Agricul-
ture for Investment. Washington, DC: World Bank.
———. 2010, Africa’s Infrastructure: A Time for Transfor-
mation. Washington, DC: World Bank.
World Bank and IMF (International Monetary Fund). 2011.
Global Monitoring Report: Improving the Odds of Achiev-
ing the MDGs. Washington, DC: World Bank.
PA R T I

Tracking Successful Growth Experiences

 CHAPTER 1

Tanzania: Growth Acceleration and

Increased Public Spending with
Macroeconomic Stability
David O. Robinson, Matthew Gaertner, and Chris Papageorgiou

n the space of a few decades, Tanzania moved from

I
TANZANIA’S GROWTH TAKE-OFF
colonialism to independence—including the union of
Sub-Saharan Africa enjoyed relatively rapid growth over the
two states (Tanganyika and Zanzibar)—to socialism to
decade leading to the 2009 global financial crisis. After many
a market-oriented developing economy. Each of these
years of poor performance, it grew faster than developed
stages involved significant change, with different economic
countries between 1995 and 2005.
institutions and economic incentives.
Tanzania experienced sustained growth acceleration over
Tanzania has emerged from this period of significant
this period (figure 1.1). This section examines when and
economic transition as one of the most rapidly growing
why this acceleration occurred.
economies in Sub-Saharan Africa. For the first time since
Most low-income countries exhibit frequent phases of
independence, it has broken out of the cycle of short-
growth, stagnation, and decline (Pritchett 2000), even where
lived accelerations in growth that has characterized many
key policies and country characteristics have been relatively
low-income countries, enjoying strong uninterrupted
stable (Easterly et al. 1993). A technique developed by Bai
growth since the mid-1990s. During the period between
and Perron (1998) can be employed to identify structural
1992/93 and 2008/09, inflation remained in single digits;
“up” and “down” breaks in economic growth trends, captur-
the debt burden fell dramatically; the level of public
ing this “stop and go” behavior of economic growth in low-
spending increased significantly, permitting expansion
income countries. A break in growth is identified as the
of public services; and international reserves rose
point after which the average growth rate is above or below
sharply (annex figures 1.A1 and 1.A2).
the previous average growth rate. Berg, Ostry, and
Achieving and sustaining rapid growth while preserving
Zettelmeyer (2008) extend this methodology to identify
macroeconomic stability represents a major achievement,
“growth spells” by modifying the procedure to determine
and significant improvements have taken place in many
sample-specific critical values when the time dimension is
aspects of development. But Tanzania remains a low-
30 years or less, the typical time horizon in low-income
income country, with per capita gross domestic product
countries growth series. This methodology was used here to
(GDP) of only $550 in 2009, and it is on track to meet only
identify episodes of growth acceleration for all countries in
about half of the Millennium Development Goals (MDGs).
Sub-Saharan Africa for which sufficiently long data series
The challenge is to harness the country’s enormous poten-
were available (table 1.1). The results reveal a large number of
tial to increase growth and thus to provide opportunities for
countries with growth accelerations in the 1990s (11 of 25).2
all to enhance living standards.1

21
 Figure 1.1 Real GDP Growth in Tanzania, 1970–2009

10

Percent 4

–2
70

75

80

85

90

95

00

05

09
19

19

19

19

19

19

20

20

20
Source: Tanzanian authorities.

making during this period was to balance the need to create

Table 1.1 Countries in Sub-Saharan Africa
Experiencing “Growth Spells” a supportive environment for growth against the need to
in the 1990s contain the potential vulnerabilities that have derailed eco-
nomic booms in the past, in Tanzania and elsewhere. Tanza-
Year of
Country acceleration
nia’s macroeconomic record has been remarkably successful
Burkina Faso 1998
during a period that has seen both external shocks (sharp
Cameroon 1994 fluctuations in global commodity prices, such as oil and
Djibouti 1998 food, and a global financial crisis) and domestic shocks
Equatorial Guinea 1994
Ghana 1997
(periodic droughts, bank failures, and governance scandals).
Liberia 1994 Navigating such shocks without major macroeconomic dis-
Mozambique 1995 ruption is a reflection of strong institutions and responsible
Namibia 1998
Rwanda 1994
policy making.
South Africa 1995 In addition, Tanzania’s move to a higher growth trajec-
Tanzania 1996 tory came following a period of substantial economic
Source: Authors’ compilation, based on data from Berg, Ostry, and reform. Three distinct phases in economic policy making
Zettelmeyer 2008. can be identified (figure 1.2).3 The first, which began at the
time of the Arusha Declaration in 1967, was the period of
Ujamaa socialism, which created a one-party system with
Tanzania’s growth take-off was spurred by several key state control of the economy and nationalization of all
factors, including the significant structural changes that major enterprises. This period ended in the mid-1980s, with
occurred as the basic institutions of a market economy—a attempts to gradually introduce key components of a
private banking system, the unification of the exchange rate, market-oriented economy.
and price liberalization—were introduced. Relative to other The pace and breadth of key reforms—both absolutely
countries that have experienced growth accelerations, and relative to the rest of the world—can be highlighted
Tanzania stands out in two respects. First, the extent of the by using newly constructed indexes of structural reforms
reforms was much broader and larger than the average of developed by the Research Department of the Interna-
other countries that saw growth accelerate. Second, there tional Monetary Fund (IMF) covering several kinds of
was a long lag between key reforms and the realization of real (trade, agriculture, and networks) and financial
the growth acceleration. The role of macroeconomic policy (domestic finance, banking, securities, and capital account)

22 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
Figure 1.2 Chronology of Transformation of Tanzania’s Economy, 1967–2009

1967–1985: Ujamaa Socialism 1986–1995: Liberalization 1996–: Macroeconomic

and Economic Decline
• State control of the economy
and state ownership of all major
enterprises
• Exchange rate and pricing
policies based on non-market
POLICIES
and Partial Reforms
• Liberalization of exchange and
trade regimes
• Liberalization of agricultural marketing
system and domestic prices
Stabilization and Structural
Reforms
• Privatization and reform of
parastatals
• Liberalization of financial sector
• Creation of market-oriented

mechanisms
• Initiation of financial system reform regulatory framework
• Devaluations and
• Trade reform, regional integration
expansionary fiscal and • Initiation of parastatal and civil service
monetary policies reforms • Reversal of fiscal dominance of
monetary policy
• Mismanagement of exogenous
shocks (terms of trade and • Fiscal consolidation
droughts)
• Sizable financial assistance from
• External trade and forex donors (debt relief, grants and
controls concessional loans)

• Low export and real GDP growth
• Deterioration of physical
infrastructure
• Loss-making state enterprises, and
large subsidies financed by directed
bank lending
OUTCOMES
• But: large segments of the
economy still dominated by public
monopolies
• Insolvency of large state-owned
banks and losses in other
parastatals
• Higher broad-based real GDP growth
• Inflation declined to single digits
• Strong growth in nontraditional exports
and turnaround in balance of payments
• Large increase in international reserves

• Composition of expenditure moved

• Budget deficits financed by printing
money
• Shortages of goods and high
inflation
• Large external imbalances,
exhaustion of foreign reserves,
and buildup of external arrears
• Rising poverty
• Persistent weaknesses in
budgetary management
• Large fiscal deficits
• Continued accumulation of arrears
and monetization of the deficits
• Elusive macroeconomic stability
• Low growth for most of the period
toward more allocations for povert-reducing
programs
• Creation of efficient, competitive
banking system
• Increased credit to the productive
sectors of the economy

Source: Nord et al. 2009.

indicators. The indexes, which date back as far as the
1970s, are summarized in annex C and in Prati, Onorato,
and Papageorgiou (2010).
There was a broad trend across much of the world
toward greater structural reform and liberalization during
the 1990s. Tanzania implemented significant reforms in the
real and financial sectors beginning in the early 1990s and
continuing into the late 1990s, reaching levels well above the
averages for Sub-Saharan Africa (an exception is the capital
account index, which lagged the region as a whole). Clear
standouts are current account and domestic financial
reforms (annex figure D1), which show extraordinary
improvements in a very short period. Figures 1.3 and 1.4
document the association between reforms (especially in
trade and finance) and growth breaks.
Various factors may explain this pattern. First, given the
nationalization of all private property that occurred in
1967, a credibility gap probably existed regarding the irre-
versibility of reforms. Second, given the physical size of
Tanzania and the limited means of communication, it took
time to convey information on the reforms throughout the
country and to translate the legislative reforms into actual
change on the ground. Third, as discussed in Mwase and
Ndulu (2008), failure to address key bottlenecks prevented
the realization of gains from other reforms. The exchange
market was unified only in 1994, for example, removing a
severe obstacle to trade, including access to needed
imports (figure 1.5).
SUSTAINING GROWTH FOLLOWING
THE INITIAL ACCELERATION
Tanzania’s economy has grown by 3.5–7.8 percent a year since
1996, averaging 6.0 percent, well above the rates for Sub-
Saharan Africa as a whole. These rates represent a substantial
increase over earlier growth rates in Tanzania and, unlike pre-
vious accelerations in growth, have been maintained over a
sustained period of time (figure 1.6).

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 23
 Figure 1.3 Indexes of Real Sector Reform in Tanzania and Other Countries Experiencing Growth Acceleration

a. Trade index
0.75

0.70

0.65

0.60 More
Index
liberalization
0.55

0.50 Less
liberalization
0.45

0.40
–5 –4 –3 –2 –1 0 1 2 3 4 5
Years before and after growth break

b. Current account restrictions index

0.90
0.80
0.70
0.60
0.50
Index

More
0.40 liberalization
0.30
Less
0.20
liberalization
0.10
0
–5 –4 –3 –2 –1 0 1 2 3 4 5
Years before and after growth break

c. Agriculture index
0.70

0.65
More
0.60 liberalization

0.55
Less
Index

0.50 liberalization

0.45

0.40

0.35

0.30
–5 –4 –3 –2 –1 0 1 2 3 4 5
Years before and after growth break

Average liberalization index Tanzania's liberalization index

Source: Authors estimates using GDP data from IMF 2010d.

Note: The number of countries used to compute each average varies across indexes, based on data availability. See annex B for a list of countries included
in sample.

24 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
Figure 1.4 Indexes of Financial Sector Reforms in Tanzania and Other Countries Experiencing Growth Acceleration

a. Domestic finance index

0.80

0.70

Index 0.60

0.50

0.40 More
liberalization
0.30 Less
liberalization
0.20
–5 –4 –3 –2 –1 0 1 2 3 4 5
Years before and after growth breaks

b. Banking subindex
1.00
0.90
0.80
0.70
0.60
Index

0.50
0.40 More
liberalization
0.30
0.20 Less
0.10 liberalization
0
–5 –4 –3 –2 –1 0 1 2 3 4 5
Years before and after growth breaks

c. Capital account index

0.54
0.52
0.50
0.48
0.46
Index

0.44 More
liberalization
0.42
0.40 Less
liberalization
0.38
0.36
0.34
–5 –4 –3 –2 –1 0 1 2 3 4 5
Years before and after growth breaks

Average liberalization index Tanzania's liberalization index

Source: Authors estimates using GDP data from IMF 2010d.

Note: The number of countries used to compute each average varies across indexes, based on data availability. See Annex B for a list of countries included
in sample.

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 25
 Figure 1.5 Official and Parallel Market Exchange Rates in Tanzania, 1970–95

600 800

700
500

Market premium (percent)

600

Shilling/U.S. dollar
400
500

300 400

300
200
200
100
100

0 0

19 4
19 5
19 6
87

19 8
19 9
19 0
91

19 2
93
94
19 0
19 1
19 2
19 3
19 4
19 5
19 6
19 7
19 8
19 9
19 0
19 1
19 2
19 3
8
8
8

8
8
9

9
7
7
7
7
7
7
7
7
7
7
8
8
8
8

19

19

19
19

Parallel market premium Parallel market rate Official rate

Source: Nord et al. 2009.

Figure 1.6 Real GDP Growth in Tanzania, 1970–2008/09

5
Percent

0
5

9
–8

–9

00

–0

–0

–0

–0
70

86

–2

05

06

07

08
19

19

96

20

20

20

20
19

Tanzania Sub-Saharan Africa

Sources: Tanzanian authorities and IMF staff estimates.

Several patterns are evident from Tanzania’s growth per- ■ The key sectors contributing to growth have been ser-
formance (figures 1.7 and 1.8 and table 1.2): vices and, to a lesser degree, industry; agriculture has
not contributed to growth. Within the service sector,
■ Acceleration has been driven by domestic demand, not particularly rapid growth has been experienced in con-
exports. Large increases in both consumption and invest- struction, telecommunications, financial services, and
ment have been recorded, in both cases reflecting signif- mining—all subsectors that have been liberalized. The
icant increases in public spending. limited contribution of agriculture—in a country where

26 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
Figure 1.7 Contributions to Real GDP Growth in Tanzania by Type of Expenditure and Production, 1999–2009

a. Contribution to real GDP growth in Tanzania by b. Contribution to real GDP growth in Tanzania by
type of expenditure sector of production
20
15
16
12
12
Percentage points

Percentage points
8 9

4
6
0
3
–4

–8 0
99

00

01

02

03

04

05

06

07

08

)
(e

99

00

01

02

03

04

05

06

07

08

)
(e
19

20

20

20

20

20

20

20

20

20

09

19

20

20

20

20

20

20

20

20

20

09
20

20
Consumption Investment Net exports Agriculture Industry Services

Sources: Tanzanian authorities and IMF staff estimates.

Figure 1.8 Contribution to Real GDP Growth in Tanzania of Labor, Capital, and Total Factor Productivity, 1995–2008

6
Percentage points

0
95

96

97

98

99

00

01

02

03

04

05

06

07

)
(e
19

19

19

19

19

20

20

20

20

20

20

20

20

08
20

TFP Capital Labor

Sources: Tanzanian authorities and IMF staff estimates.

about three-quarters of the population resides in rural ■ The acceleration in growth can be traced to a combi-
areas, where poverty is concentrated—is a concern. It also nation of higher investment and increases in produc-
represents an opportunity, because international experi- tivity. Growth decomposition is always subject to
ence suggests that the sector can respond rapidly if the interpretation, particularly regarding the estimation of
right incentives and supporting infrastructure are put the capital stock (and therefore the contribution of
in place. investment and the interpretation of the residual as

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 27

Table 1.2 Factor Contributions to Real GDP
Growth in Tanzania, 1986–2008
(percent)
Item 1986–90 1991–95 1996–2000 2001–08
Real GDP growth 5.3 1.8 4.3
Labor force 2.2 2.5 1.7
Capital 0.9 1.3 0.3
Total factor
productivity 2.2 –2.0 2.3
Source: Tanzanian authorities and IMF staff calculations.
productivity growth). That said, it is clear that the
acceleration in Tanzania was not driven by greater use
of labor.4 Studies of the agricultural sector during this
period show little or no improvement in yields for the
sector as a whole, with increased output coming from
an increase in land under cultivation.
Although exports did not lead the growth acceleration—
indeed at just 25 percent of GDP in 2008/09, they remain at
a low level—the composition of exports has seen significant
changes (figure 1.9). Traditional exports (cotton, coffee, and
tea) have declined significantly in importance, partly reflect-
ing the reorientation of the sector toward meeting the
consumption needs of a rapidly growing local population.
At the same time, gold exports went from zero in 1999 to
$1.4 billion (nearly 40 percent of export receipts) in 2009,
making Tanzania the fourth-largest gold exporter in Africa
after South Africa ($6.3 billion), Ghana ($2.6 billion), and
Mali ($2.0 billion). The service sector—tourism and other
services, such as transport—has also grown at a steady rate.
The external environment
Tanzania’s growth acceleration coincided with a period of
both unprecedented expansion in the global economy and
significant volatility in commodity prices and capital flows
(private and official, including both aid and debt relief) that
have posed challenges for macroeconomic management.
Given its low level of exports, Tanzania has been insulated
from fluctuations in prices and demand for exports and the
boom-bust cycles such fluctuations have created in other
commodity-dependent countries. However, significant
impacts have been felt on the import side. The sharp
increase in oil prices during the 2000s resulted in an oil
import bill of $1.7 billion in 2007/08, almost 9 percent of
GDP. Recent spikes in the price of fertilizer, most of which
is imported, have constrained the government’s ability to
28 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
stimulate agricultural production. Disruptions in the global
rice market in 2008 hit Zanzibar hard, contributing to infla-
tion that reached 27 percent a year in September 2008
before declining to 5 percent a year later.
Fairly broad restrictions on capital account transactions
7.1
1.7 remain in place—preventing, for example, foreign portfolio
1.9 inflows to the domestic government securities markets or
private capital outflows into neighboring securities markets.
3.5
Private capital inflows are nevertheless becoming an
increasingly important source of financing for investment.
Foreign direct investment, most of it in the mining sector,
averaged 3.5 percent of GDP during 1996–2008. Before the
global financial crisis, loan syndications were being used to
finance investment, mostly in the telecom sector.
Official capital inflows have been important sources of
financing for Tanzania since independence. Official devel-
opment assistance peaked on average during 1986–92 (it
represented almost 25 percent of GDP in 1992) before
declining through the 1990s and then rising again, to about
17 percent of GDP in 2007. The component of aid that goes
through the budget has been rising, albeit with significant
year-on-year variability (table 1.3).
In addition to the absolute level of aid inflows, two other
aspects of aid have had macroeconomic impacts. First, the
composition of aid inflows has shifted. In line with the gov-
ernment’s stated preferred modality, aid has increasingly
been provided in the form of general budget support—cash
that goes directly into the budget—rather than project sup-
port. Such aid gives the government more control over the
use of resources provided by development partners and can
be used to better meet national rather than externally driven
(foreign) priorities. However, the use of general budget sup-
port also places a heavier burden on government processes
to ensure the effective allocation of budgetary resources and
maintain a broad dialogue with development partners on
results and use of funds to ensure continued access to aid.
Its use also complicates macroeconomic management, mak-
ing it critical to ensure that the spending of aid does not
induce undesirable macroeconomic consequences, such as a
sharp appreciation of the exchange rate. An early evaluation
of the impact of general budget support in Tanzania noted
that it facilitated the government’s ability to implement
policies to which it was committed without discernable
adverse macroeconomic consequences, but that it was less
effective in securing commitment to policies (Lawson
et al. 2005).
Significant aid has been provided in the form of debt
relief. Tanzania qualified for the Heavily Indebted Poor
Countries Initiative in 2001, receiving debt relief of about
Figure 1.9 Composition of Exports of Goods and Services in Tanzania, 1996/97–2008/09

100

80

60
Percent

40

20

0
7

9
/9

/9

/9

/0

/0

/0

/0

/0

/0

/0

/0

/0

/0
96

97

98

99

00

01

02

03

04

05

06

07

08
19

19

19

19

20

20

20

20

20

20

20

20

20
Gold Traditional Other goods Other services Tourism

Source: Tanzanian authorities.

Table 1.3 Gross External Program and Project Assistance in Tanzania, 1996/97–2007/08
Program assistancea Project assistance
Year Grants Loans Grants Loans Debt service relief Total assistance
1996/97 1.7 0.9 3.0 3.2 0.0 8.7
1997/98 1.7 1.3 2.6 2.2 0.0 7.8
1998/99 1.7 1.2 2.5 1.9 0.0 7.3
1999/00 1.7 0.7 2.8 2.8 0.0 8.1
2000/01 1.6 0.5 3.2 2.1 0.4 7.8
2001/02 2.4 0.9 2.8 1.1 0.6 7.8
2002/03 3.2 1.3 2.3 1.1 0.6 8.5
2003/04 3.4 1.7 2.6 2.0 0.6 10.2
2004/05 4.2 1.3 2.7 2.1 0.5 10.7
2005/06 3.5 2.0 2.0 1.9 0.7 10.1
2006/07 3.3 1.7 1.2 2.2 1.2 9.7
2007/08 3.5 2.4 3.0 2.2 0.9 12.1

Source: Tanzanian authorities and IMF staff estimates.

a. Includes both general budget support and basket funds.

$3 billion. In 2006 the IMF, the World Bank, and the African
Development Bank implemented the Multilateral Debt
Reduction Initiative, providing debt relief of an additional
$3.5 billion.
Debt relief has substantially reduced the debt burden in
Tanzania, freeing up budgetary resources for alternate uses.
It has also left the country with a low level of debt—in
absolute and especially net present value terms, given the
highly concessional nature of much of the outstanding
stock—generating a low risk of debt distress in conventional
debt sustainability analyses (see, for example, the joint
Bank/Fund debt sustainability analysis in IMF 2010c) and
making Tanzania a potentially attractive destination for
capital inflows. Cognizant of the difficulties that arose from
the very high debt burden, in 2004 lawmakers introduced a
formal minimum concessionality requirement in external
borrowing by the government, initially established at
50 percent but subsequently reduced to 35 percent.
Fiscal policy
The most striking feature of fiscal policy between 1992/93
and 2008/09 was the expansion in public spending. After
falling to 15–17 percent of GDP in the mid-1990s as inflation

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 29
 was brought under control, spending increased sharply in
2001/02, reaching more than 25 percent of GDP in 2008/09
(even higher figures were budgeted for 2009/10 and
2010/11) (figure 1.10). Government spending per capita
increased from about $40 in 1998/99 to $150 in 2008/09.
How was this spending financed without endangering
price stability or increasing debt sustainability? Where did
the money go? Expenditures are financed by increasing
revenues, obtaining financing from abroad (as loans or
grants), or borrowing domestically. Tanzania has used all
three sources (figure 1.11).
Revenue performance has been striking. After an initial
dip in the second half of the 1990s, revenues increased
strongly, from just 10 percent of GDP in 1999/2000 to
almost 16 percent in 2008/09. The improvement reflects a
combination of tax policy and administration.
Tax policy has focused on reorienting the tax system to
reflect the changing economic structure, as parastatals were
gradually privatized; trade (internal and external) was liber-
alized; and a private sector, often informal, emerged. Key
changes included the following:
■ A value added tax (VAT) was introduced in July 1998,
providing about one-third of government revenues. The
rate was reduced to 16 percent in July 2009, bringing it
into line with most other East African countries.
■ Income taxes, which also provide about one-third of gov-
ernment revenues, were progressively modernized, with
the maximum rates reduced to 30 percent, down from
70 percent rate in 1991.
Figure 1.10 Government Expenditure in Tanzania as
Percent of GDP, 1992/93–2008/09
30
25
20
Percent
■ The tax system was simplified, through restrictions on
the range of taxes subnational authorities can impose
and the repeal of numerous low-yielding taxes.
■ Relative stability was achieved in key tax rates. Following
initial reforms, the rates of the major taxes were left
largely unchanged, in part because they were linked to
regional tax harmonization initiatives.5
Improvements in tax administration have been no less
important. The first step in strengthening administration
came with the creation of the Tanzania Revenue Authority,
in 1998/99. The Tanzania Revenue Authority has been able
to progressively modernize tax collections, by introducing
self-assessment procedures and electronic filing and pay-
ments and by carefully allocating its own human resources
into high-return areas. Indeed the Large Taxpayer Depart-
ment introduced in 2003 now collects almost three-quarters
of domestic taxes, with the number of enterprises covered
rising from 98 to more than 370.
There is still scope for improving tax policy. In addition
to its revenue-raising role, tax policy has been used as a
component of the broader development strategy to pro-
vide incentives for specific economic activities in an
attempt to attract investment. Exemptions, such as tax hol-
idays, are provided for qualified investments and for com-
panies established in designated export-processing zones
or special economic zones. A special regime was also cre-
ated for mining activities, based on a tax/royalty regime
but with additional relief from VAT and other taxes, such
as the fuel levy. The efficacy of such incentive regimes for
attracting investment is the subject of a fierce debate and
beyond the scope of this chapter. Recent budgets have also
introduced VAT exemptions for specific industries or
products, thereby eroding the tax base and complicating
tax administration.
Reliance on domestic financing—initially largely from
the central bank, later from commercial banks and social
security funds—has decreased significantly, in part out of a
desire to contain the rate of monetary growth to bring down
inflation. For several years in the mid-2000s, budgets were

15 anchored on a zero net domestic financing target, which

provided continued support for inflation stabilization,
10
eased interest rate pressures, and increased room for banks
5 to lend to the private sector. Where there were shortfalls in
foreign financing, limited resort was made to domestic
0 financing in order to avoid disruptions in expenditure
3

programs.6
/9

/9

/9

/9

/0

/0

/0

/0

/0
92

94

96

98

00

02

04

06

08
19

19

19

19

20

20

20

20

20

Even with this dramatic improvement in revenues,

Source: Tanzanian authorities. donor dependency has not declined. Government revenues

30 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
Figure 1.11 Sources of Fiscal Expenditure in Tanzania, 1992/93–2008/09

a. Sources of fiscal expenditure as b. Sources of fiscal expenditure as

percentage of GDP percentage of expenditure
110
30
100
25 90
80
20 70
Percent

Percent
60
15
50
10 40
30
5 20
10
0
0
–5 –10

9
3

/9

/9

/9

/9

/0

/0

/0

/0

/0
/9

/9

/9

/9

/0

/0

/0

/0

/0

92

94

96

98

00

02

04

06

08
92

94

96

98

00

02

04

06

08

19

19

19

19

20

20

20

20

20
19

19

19

19

20

20

20

20

20
Revenue Grants Foreign Domestic

Source: Tanzanian authorities and IMF staff calculations.

still cover only about 50–60 percent of government
expenditures—less than the level of recurrent spending
(see figure 1.11).
Weak domestic revenue perpetuates reliance on foreign
support, potentially adding both uncertainty and volatility
to the government’s ability to deliver its programs. Alterna-
tively, the low share of government revenues to expenditures
can be interpreted as an effort by the government to exploit
the availability of cheap financing sources to expand deliv-
ery of public services beyond the level possible from domes-
tic resources alone. The key is to ensure that efforts are
under way to build a broad tax base capable of sustaining
the desired level of public services in the event that donor
funds diminish over time, that these efforts are not reduced
by the presence of aid inflows, and that the additional public
spending is used wisely.
The higher spending has appropriately been focused on
priorities related to the National Strategy for Growth and
Reduction of Poverty (known by its Swahili acronym,
MKUKUTA). The 2009/10 budget allocates 71 percent of
expenditures to MKUKUTA priorities. At the sectoral
level, the two sectors receiving the largest budget alloca-
tions are education and health, which together account for
about 30 percent of budgetary spending. The impact of the
spending is difficult to gauge. In education, for example,
substantial investment in construction of new classrooms
has supported an expansion in primary school enrollment
to about 95 percent, up from about 60 percent a decade
earlier. But issues remain with unfilled vacancies for teach-
ers, particularly in rural communities; uneven financial
transfers; and concerns about the quality of education,
with pass rates in standardized examinations declining. In
the health sector, a recent value-for-money study identi-
fied sharp regional differences in activity levels and per-
formance, with no linkage to resource allocations
(National Audit Office 2010).
Monetary and exchange rate policies
The primary objective of the Bank of Tanzania is to main-
tain domestic price stability conducive to balanced and
sustainable growth of the national economy. A secondary
objective is to support the integrity of the financial system.
With inflation reduced to single digits and a financial sys-
tem that has grown rapidly in recent years, the Bank of
Tanzania appears on track to meet both objectives.
It has not been an easy transition. In the late 1980s Tanza-
nia had one of the least developed financial systems in the
world. Credit was allocated centrally; all of the major finan-
cial institutions were state owned; there were no capital or
money markets, so the Bank of Tanzania simply printed
money to finance the fiscal deficit and provide liquidity to
insolvent banks; and foreign exchange was rationed, with a
substantial spread between the official and black market rates.

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 31
 Interest rates and foreign exchange markets were pro-
gressively liberalized starting in 1991, but the initial distor-
tions—the legacy constraints—took time to resolve and
served to shape economic policy options for most of the
decade. In particular, with minimal foreign exchange
reserves, central bank interventions in the market were ini-
tially targeted at accumulating reserves while smoothing
seasonal volatility in the exchange rate. As the government
was able to rein in its domestic financing needs and donors
increasingly moved to providing budget support, the avail-
ability of foreign exchange increased, allowing the Bank of
Tanzania to accumulate reserves, reducing upward pressure
on the exchange rate (figure 1.12). It did so at the cost of
assigning the burden of liquidity management to Treasury
bill sales, however, resulting in rising domestic debt and
upward pressure on nominal interest rates.
This unbalanced monetary policy mix, with high yields
on government securities, low inflation, and exchange rate
stability, proved difficult to sustain, however, particularly
given the shallow domestic financial market with a small
number of key players (Abbas, Ali, and Sobolev 2008). In late
2007 the central bank issued a press release indicating that its
foreign exchange operations were limited to smoothing
short-term volatility and that the exchange rate could appre-
ciate. Only thereafter did markets become more balanced.
Private banks
Private banks were formally permitted in 1992, but none
began operations until 1994. All of the large state-owned
banks were privatized, but the process was lengthy and the
former state-owned banks—CRDB (privatized in 1996),
NBC (privatized in 2000), and NMB (privatized in 2005)—
continue to dominate the financial system. Indeed, only
since about 2000 has the private sector been able to access
credit from the banking system.
Credit has grown rapidly since then (figure 1.13). Entry of
new banks has accelerated in recent years, with 41 banks
licensed as of the end of June 2010. Access to financial ser-
vices remains limited, however—only 11 percent of adult
Tanzanians held a bank account at end-2009—with financial
services in rural areas increasingly shifting to mobile tele-
phone systems.
Financial sector supervision has evolved along with
the changing role of Tanzania’s financial institutions. The
government and the Bank of Tanzania developed a risk-
based supervisory framework that provides sufficient
authority to the body responsible for supervising and
regulating the banking system. A bank intervention and
resolution framework is in place, supported by a deposit
insurance scheme. These institutions facilitated the clo-
sure of four private banking institutions, including two
foreign-owned, without major disruptions to the banking
system and with full compensation for individual deposi-
tors.7 The legislative foundation for supervision of the
broader financial system—chiefly pension funds and
insurance companies—has been put in place, but imple-
mentation needs to be enhanced to guard against the
buildup of potentially large fiscal liabilities.

Figure 1.12 Foreign Exchange Reserves in Tanzania, 1992/93–2008/09

3,500

3,000

2,500
Millions of dollars

2,000

1,500

1,000

500

0
19 94

19 95

19 96

19 97

19 98

19 99

20 00

20 01

20 02

20 03

20 04

20 05

20 06

20 07

20 08
9
/0
/

/
93

94

95

96

97

98

99

00

01

02

03

04

05

06

07

08
19

Source: Tanzanian authorities.

32 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
Figure 1.13 Private Sector Credit as Percent of GDP in Tanzania, 1992/93–2008/09

18

16

14

12

Percent
10

0
19 6

19 7

19 8

19 9

20 0

20 1

20 2

20 3

20 4

20 5

20 6

20 7
8

9
/9

/9

/9

/9

/0

/0

/0

/0

/0

/0

/0

/0

/0

/0
95

96

97

98

99

00

01

02

03

04

05

06

07

08
19

20
Source: Tanzanian authorities and IMF staff calculations.

USING THE NEWFOUND POLICY SPACE: Despite initial concerns that contributed to a request
COUNTERCYCLICAL POLICY RESPONSE for financing from the IMF under the exogenous shocks
TO THE GLOBAL FINANCIAL CRISIS facility, Tanzania has managed to contain the impact of the
global financial crisis through a macroeconomic policy
The degree to which Tanzania, and many other countries in
response supported by interventions at the microeco-
Africa, have changed in terms of both economic structures
nomic level, all laid out in an economic recovery plan
and macroeconomic policy space was clearly demonstrated
launched in June 2009. At the macroeconomic level, both
during the global financial crisis. Past global slowdowns
fiscal and monetary policies were eased to help sustain the
have been very difficult for Africa: as documented in IMF
growth momentum. On the fiscal side, easing of more than
(2010a), in previous global slowdowns, Africa suffered both
2 percent of GDP was envisaged, including a reduction in
a deeper slowdown and a more gradual recovery. This crisis
the VAT rate from 20 percent to 18 percent, partial gov-
has been different, with Sub-Saharan Africa as a whole
ernment guarantees for loan restructurings in troubled
seeing a large but short-lived dip in economic growth
sectors, expanded agricultural input subsidies, expanded
(figure 1.14).
investments in energy and road sectors, and temporary
Another key difference is that, for the first time, many
exemptions from royalties for tanzanite and diamond
countries in Sub-Saharan Africa were in a position to pur-
miners. The interventions were specifically targeted to sec-
sue countercyclical policies in the face of an adverse shock.
tors expected to be hardest hit by the crisis. The fiscal eas-
In response to previous shocks, African countries generally
ing was financed in part by development partners, several
pursued procyclical macroeconomic policies, largely out of
of which were able to advance funding from future years,
lack of choice: high debt burdens, limited sources of financ-
but also in part from domestic sources, with a modest
ing, and low levels of international reserves provided few
increase in the volume of sales of Treasury bills and excep-
options for easing fiscal or monetary policy in the face of a
tional direct credits from the central bank. As elsewhere,
growth slowdown. In addition, changes in the structure of
commercial banks became more risk averse, with credit
the global economy—specifically the increased role of
growth to the private sector falling sharply and high
China and India, both of which are key trading partners for
demand for government obligations serving to push inter-
much of Sub-Saharan Africa and which were able to contain
est rates to record lows. Although the crisis is not yet over,
the damage to their economies—mitigated the external
preliminary estimates for 2009 point to growth slowing to
demand shock experienced by countries in Sub-Saharan
6.0, rising to 6.5 percent in 2010.
Africa from the global financial crisis.

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 33
 Figure 1.14 Real GDP Growth in Sub-Saharan Africa and the World, 1992/93–2008/09

a. Real GDP growth, average of past three

economic cycles, 1992/93–2008/09 b. Real GDP growth, current cycle 2005–2013
6 8

5
6

4
4
Percent

Percent
3
2
2

1 0

0 –2
t–4 t–3 t–2 t–1 t t+1 t+2 t+3 t+4

05

06

07

08

09

10

11

12

13
20

20

20

20

20

20

20

20

20
World Sub-Saharan Africa

Source: IMF 2010d.

Note: t = trough in the cycle.

WHAT’S NEXT?
Tanzania has managed to achieve and sustain rapid growth
while preserving macroeconomic stability for the past
15 years. Despite this achievement, it is on track to achieve
only about half of the MDGs (table 1.4) (United Republic of
Tanzania 2009b). What needs to be done in the years ahead?
What macroeconomic policies would be beneficial?
Accelerating pro-poor growth
Constraints to growth in Tanzania are similar to those seen
in most countries: inadequate infrastructure, regulatory
bottlenecks, skill shortages, and deficiencies in the legal
environment. These weaknesses constrain the realization of
income opportunities at all levels—from the smallholder
farmer trying to get surplus crops to market to the large for-
eign investor.
The past few years have seen little progress in several
indicators of the business environment. Tanzania’s rank in
the Doing Business Index has been slipping, reaching 131 in
2010, as other countries have taken more determined strides
to make their economies more business friendly. Tanzania
scores particularly poorly on basic bureaucratic require-
ments involving numbers of procedures and permits and
the time spent to receive them. The importance of address-
ing this slowdown in these broader areas of structural
reform is reinforced by a key result of the literature on
growth spells: although macroeconomic and financial
reforms can drive a growth up-break, they are not sufficient
in isolation to sustain one (Hausmann, Pritchett, and
Rodrik 2005).
Infrastructure is a constraint across Africa. But Tanzania
lags even other countries in Africa (table 1.5).
Where will growth come from? Studies of potential
growth drivers for Tanzania (Mbelle et al. 2010, for example)
highlight a combination of agriculture, tourism, transport,
and mining, with a focus on developing manufacturing and
value addition activities. The identified sectors largely
reflect physical and geographical endowments, with existing
activities in these sectors fairly nascent—only a small por-
tion of Tanzania’s natural resource endowments are being
exploited—and often characterized by low productivity. In
many countries developing agriculture has been key to
ensuring sustained progress toward the MDGs (World Bank
2005). Tanzania has decided to emphasize agriculture, hav-
ing launched the Kilimo Kwanza (Agriculture First) cam-
paign designed to accelerate reforms in the sector, including
expanding input subsidy programs and significant invest-
ments in irrigation.
Meeting macroeconomic challenges
Addressing the infrastructure deficit will be expensive and
will likely require funding beyond that likely to be available
on the highly concessional terms at which Tanzania has
been borrowing. The additional funds could come from
many potential sources, including donor-financed projects,

34 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
 Table 1.4 Projected Progress toward Selected MDG Targets by Tanzania
Target 1990 2000 2008 2015 Likely to achieve target?
Population below basic needs poverty line (percent) 39.0 36.0 33.6 19.5 No
Under-five underweight (percent) 28.5 29.5 22.0 14.4 No
Under-five stunted (percent) 46.6 44.4 38.0 23.3 No
Primary school net enrollment rate 54.2 58.7 97.2 100.0 Yes
Under-five mortality rate (per 1,000 live births) 191.0 153.0 112.0 64.0 Yes
Infant mortality rate (per 1,000 live births) 115.0 99.0 68.0 38.0 Yes
Maternal mortality rate (per 100,000 live births) 529.0 — 578.0 133.0 No
Births attended by skilled health personnel (percent) 43.9 35.8 63.0 90.0 No
HIV prevalence, 15–24 (percent) 6.0 — 2.5 < 6.0 Yes
Access to potable water (percent of rural population) 51.0 42.0 57.1 74.0 No
Access to potable water (percent of urban population) 68.0 85.0 83.0 84.0 Yes

Source: United Republic of Tanzania 2009a.

Note: — Not available.

Table 1.5 Infrastructure Indicators in Selected Countries in Sub-Saharan Africa

Households Mobile Households Access to Access to
with fixed phones with electricity Roads improved improved
telephone (subscribers (percent (km per sanitation water source
Country or (percent of per 100 connected to 1,000 km2 (percent of the (percent of the
country group households)a people)b network) a of land)c population)a population)a
Ghana 8 32 44 187 10 80
Kenya 12 30 13 111 42 57
Rwanda 1 7 5 568 23 65
South Africa 27 92 63 300 59 93
Tanzania 10 21 11 62 33 55
Uganda 3 18 8 385 33 64
Zambia 4 21 20 50 52 58
Sub-Saharan Africa 7 16 29 — 31 58
Sub-Saharan Africa low incomed 6 19 26 — 15 25
Sub-Saharan Africa middle incomed 19 36 55 — 41 66

Source: World Bank Africa Infrastructure Country Diagnostic Database and World Bank various years.
Note: — Not available.
a. Data are from Demographic and Health Surveys, latest available year for 2001–08.
b. Data are for 2006.
c. Data are for latest available year 2001–08.
d. Income groups are based on World Bank classification.

local or international bonds, and public-private partner-
ships (Ter-Minassian, Hughes, and Hajdenberg 2008).
Countries in Sub-Saharan Africa are increasingly accessing
international capital markets, in several cases in the context
of IMF-supported programs (Redifer 2010). Whatever the
source of financing, it will be critical to ensure strong gov-
ernment processes for project selection, planning, and
implementation and effective debt management to maxi-
mize the probability of a strong economic return while
containing macroeconomic risks. Additional capacity to
monitor risks in public-private partnerships will also be
important to avoid the emergence of potentially significant
contingent liabilities (the track record of such partnerships
in Tanzania and more generally in Sub-Saharan Africa is
littered with examples of projects in which the host gov-
ernment ended up either simply taking over the project or
having to renegotiate in order to provide better terms for
the private sector partner) (Gratwick and Eberhard 2008).
Fiscal pressures—not just from infrastructure spending
but also from growing demand for public services as a result
of high population growth—could be eased by mobilizing
additional domestic resources. Estimates of tax potential
suggest that revenues could reach 21 percent of GDP, almost
5 percentage points more than current collections (IMF
2010c).8 A big part of the revenue loss stems from exemp-
tions, which amounted to 30 percent of tax collections
(3.5 percent of GDP) in 2007/08, according to estimates by
the Tanzania Revenue Authority, making Tanzania’s VAT

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 35
 one of the least efficient in the region (table 1.6). A possibly
substantial upside is natural resource revenues: the mining
sector is expanding rapidly and, as exemptions are reined in
and those for the more established mines expire, the tax take
can be expected to increase significantly. Close cooperation
and transparency in the granting of exemptions will also
need to be tackled in the context of regional integration
efforts, in order to avoid a race to the bottom, with poten-
tial investors able to induce neighboring governments to
compete against one another on the physical location of
activities, resulting in the erosion of the tax base.
The exchange rate may also pose challenges. To date, the
flexibility in the exchange rate, the apparent absence of
resource capacity constraints in the economy, and produc-
tivity growth have all served to avoid the emergence of a
misalignment of the exchange rate (see Hobdari 2008).
Going forward, in the event of the planned scaling up of
both public spending and private sector investment, emerg-
ing capacity constraints will need to be monitored carefully,
and it will be important to ensure that public spending is
channeled into productivity-enhancing investment to sup-
port the competitiveness of the economy. Similarly, the move
to a common currency for the East African Community—
which could generate additional growth opportunities—
will need to address the issue of how individual national
economies will adjust to localized shocks once they relin-
quish the ability to adjust through the exchange rate.
Limiting vulnerabilities but preparing for the worst
Even with the best macroeconomic policies and monitoring
systems, bad outcomes can happen. Timely policy responses
are often critical for containing the impact of bad outcomes.
Being able to craft them requires having appropriate sys-
tems in place before the shock occurs.
Table 1.6 VAT Revenue and Revenue Productivity in Selected Countries in
Sub-Saharan Africa, 2009/10
Current standard VAT
Country rate (percent)
Kenya 16.0
Malawi 16.5
Rwanda 18.0
South Africa 14.0
Tanzania 18.0
Uganda 18.0
Unweighted average 16.8
Source: IMF VAT database and staff projections.
a. Projection for 2009/10.
b. Total VAT revenue, as percent of GDP, divided by the standard rate.
36 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
In the financial sector, Tanzania’s basic crisis resolution
framework, together with a deposit insurance scheme, was
able to handle failures of individual banks. That framework
should be continuously reviewed to ensure that it can meet
potential risks; there is room to further strengthen banking
supervision, including risk analysis and enforcement of
prudential requirements, while addressing staffing con-
straints (IMF 2010b).
The absence of effective oversight of pension funds is a
critical deficiency in the current regulatory framework. Lit-
tle information is available on the financial position of pen-
sion funds, their role in capital markets, and potentially
large fiscal liabilities.
With the increasing integration of the financial system of
the East African Community, there is also a need to ensure
that a coordinated policy response or intervention can be
orchestrated in the event of, say, pressure on a systemic
institution. The initial difficulty in agreeing on a coordi-
nated response was perhaps one of the key features of the
global financial crisis.
More generally, a well-defined social safety net can help
protect the most vulnerable in the event of economic down-
turns. All of these responses require policy flexibility.
Retaining the policy buffers that were used to such good
effect during the global financial crisis is critical for the con-
duct of countercyclical policy and the ability to prevent an
adverse shock from derailing the economy’s growth
momentum.
CONCLUSION
Tanzania has seen unprecedented sustained growth accel-
eration since 1996. Several major (real and financial)
reforms played a pivotal role in the take-off. Maintaining
macroeconomic stability was a critical component in
VAT revenue Revenue productivity
(percent of GDP)a (percent)b
6.0 37.5
6.0 37.0
6.0 33.3
5.7 40.5
4.6 25.5
4.0 22.2
5.4 32.7
sustaining growth. Navigating the post–1996 period has Is another 15 years of uninterrupted growth likely?
involved the development of new policy instruments, as Untapped growth potential clearly remains, and the policy
the nature of the economy changed and legacy constraints space that has been developed over the past decade or so can
were eased. provide needed support.

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 37
 ANNEX 1.A MACROECONOMIC PERFORMANCE

Figure 1.A1 Macroeconomic Performance in Tanzania and Sub-Saharan Africa, 1970–2009

a. Real GDP growth b. Inflation
8 35
7 30
6
25
5
Percent

Percent
20
4
15
3
10
2
1 5

0 0
5

9
–8

–9

–0

–0

–0

–0

–0

–8

–9

–0

–0

–0

–0

–0
70

86

96

05

06

07

08

70

86

96

05

06

07

08
19

19

19

20

20

20

20

19

19

19

20

20

20

20
c. GDP per capita
700

600

500

400
US$

300

200

100

0
5

9
–8

–9

–0

–0

–0

–0

–0
70

86

96

05

06

07

08
19

19

19

20

20

20

20

Tanzania Sub-Saharan Africa

38 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
Figure 1.A2 External Sustainability Indicators in Tanzania, 1970–2009
a. Public sector external debt b. Public sector external debt
120 40
35
100
30

Percent of exports
Percent of GDP

80
25
60 20
15
40
10
20
5
0 0
5

9
–8

–9

–0

–0

–0

–0

–0

–8

–9

–0

–0

–0

–0

–0
70

86

96

05

06

07

08

70

86

96

05

06

07

08
19

19

19

20

20

20

20

19

19

19

20

20

20

20
c. Foreign reserves
6

5
Months of imports

0
5

9
–8

–9

–0

–0

–0

–0

–0
70

86

96

05

06

07

08
19

19

19

20

20

20

20

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 39
 ANNEX 1.B LIST OF COUNTRIES IN THE SAMPLE

Table 1.B1 List of Countries in the Sample

Low income Middle income High income
Bangladesh Albania Australia
Burkina Faso Algeria Austria
Côte d’Ivoire Argentina Belgium
Ethiopia Azerbaijan Canada
Ghana Belarus Czech Republic
India Bolivia Denmark
Kenya Brazil Estonia
Madagascar Bulgaria Finland
Mozambique Cameroon France
Nepal Chile Germany
Nigeria China Greece
Pakistan Colombia Hong Kong SAR, China
Senegal Costa Rica Ireland
Tanzania Dominican Republic Israel
Uganda Ecuador Italy
Uzbekistan Egypt, Arab Rep. of Japan
Vietnam El Salvador Korea, Rep. of
Zimbabwe Georgia Netherlands
Guatemala New Zealand
Hungary Norway
Indonesia Portugal
Jamaica Singapore
Jordan Spain
Kazakhstan Sweden
Latvia Switzerland
Lithuania Taiwan, China
Malaysia United Kingdom
Mexico United States
Morocco
Nicaragua
Paraguay
Peru
Philippines
Poland
Romania
Russian Federation
South Africa
Sri Lanka
Thailand
Tunisia
Turkey
Ukraine
Uruguay
Venezuela, R. B. de

Source: World Bank classification.

40 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
 ANNEX 1.C MEASURES OF REFORM

Table 1.C1 Measures of Reform

Coverage

MIN # of MAX # of
countries in countries in
Measure Description Source Start year End year any year any year
Real Indices
Trade Openness
Tariff Rates Average tariff rates, with missing values extrapolated Various sources, including IMF, 1960 2005 47 142
using implicit weighted tariff rates. Index World Bank, WTO, UN, and
normalized to be between zero and unity: zero the academic literature
means the tariff rates are 60 percent or higher, (particularly Clemens and
while unity means the tariff rates are zero. Willamson, 2004)
Current-Account An indicator of how compliant a government is with Quinn (1997), and Quinn and 1960 2005 50 65
Restrictions its obligations under the IMF’s Article VIII to free Toyoda (2007; forthcoming).
from government restriction the proceeds from
international trade in goods and services. The
index represents the sum of two sub-components,
dealing with restrictions on trade in visibles, as
well as in invisibles (financial and other services).
It distinguishes between restrictions on residents
(receipts for exports) and on non-residents
(payments for imports). Although the index
measures restrictions on the proceeds from
transactions, rather than on the underlying
transactions, many countries in practice use
restrictions on trade proceeds as a type of trade
restriction. The index is scored between zero and
8 in half-integer units, with 8 indicating full
compliance.
Product Markets
Telecom and Electricity Simple average of the electricity and telecom Based on legislation and other 1960 2003 106 108
Industries markets sub-indices, which are constructed, in official documents.
turn, from scores along three dimensions. For
electricity, they capture: (i) the degree of
unbundling of generation, transmission, and
distribution; (ii) whether a regulator other
than government has been established; and
(iii) whether the wholesale market has been
liberalized. For telecom, they capture:
(i) the degree of competition in local services;
(ii) whether a regulator other than government
has been established; and (iii) the degree of
liberalization of interconnection changes.
Indices are coded with values ranging from zero
(not liberalized) to two (completely liberalized).

(continued next page)

41
42

Table 1.C1 (continued)

Coverage

MIN # of MAX # of
countries in countries in
Measure Description Source Start Year End Year any year any year
Real Indices
Agriculture Given that developing countries constitute most of Based on legislation and other 1960 2003 96 104
our sample, the degree of regulation in agriculture, official documents.
which continues to account for a large part
of many of these economies, is an essential aspect
of product market competition. Index aims to
capture intervention in the market for the main
agricultural export commodity in each country.
As data limitations preclude coding separate
dimensions of intervention, the index provides a
summary measure of intervention. Each
country-year pair is assigned one of four degrees
of intervention: (i) maximum (public monopoly or
monopsony in production, transportation, or
marketing); (ii) high (administered prices);
(iii) moderate (public ownership in relevant
producers, concession requirements); and
(iv) no intervention.

Financial Indices
Capital Account Qualitative indicators of restrictions on financial Abiad and others (2008), which 1973 2005 72 91
Opennes: Aggregate credits and personal capital transactions of residents follows the methodology in
and financial credits to nonresidents, as well as the Abiad and Mody (2005). The
use of multiple exchange rates. Index coded from original sources are mostly
zero (fully repressed) to three (fully liberalized). various IMF reports and working
papers, but also central bank
websites, etc.
Resident/nonresident-specific
indices are based on Quinn
(1997), and Quinn and
Toyoda (2007).
 Capital Account Measures the extent to which residents
Openness: Residents (nonresidents) are free from legal restrictions to
(nonresidents) only move capital into and out of a country.
Domestic Financial The index of domestic financial liberalization is an
Liberalization average of six subindices. Five of them relate to
banking: (i) interest rate controls, such as floors or
ceilings; (ii) credit controls, such as directed credit,
and subsidized lending; (iii) competition restrictions,
such as limits on branches and entry barriers in the
banking sector, including licensing requirements or
limits on foreign banks; (iv) the degree of state
ownership; and (v) the quality of banking supervision
and regulation, including power of independence of
bank supervisors, adoption of a Basel I capital
adequacy ratio, and framework for bank inspections.
The sixth subindex refers to the regulation of
securities markets, including policies to encourage the
development of bond and equity markets, and to
permit access of the domestic stock market to
foreigners. The subindices are aggregated with equal
weights. Each subindex is coded from zero (fully
repressed) to three (fully liberalized).

Source: Prati, Onorato, and Papageorgiou (2010).

43
44

Table 1.C2 Description of Reform Indices

Coverage
MIN # of MAX # of
countries in countries in
Reform indices Description Source Start year End year any year any year
Financial Sector
Domestic Financial The index of domestic financial liberalization is an Abiad and others (2008), 1973 2005 72 91
Sector Liberalization average of six subindices. Five of them relate to following the methodology in
banking: (i) interest rate controls, such as floors Abiad and Mody (2005), based
or ceilings; (ii) credit controls, such as directed on various IMF reports and
credit and subsidized lending; (iii) competition working papers, central bank
restrictions, such as limits on branches and entry websites, and others.
barriers in the banking sector, including licensing
requirements or limits on foreign banks; (iv) the
degree of state ownership; and (v) the quality of
banking supervision and regulation, including
power of independence of bank supervisors,
adoption of Basel capital standards, and a
framework for bank inspections.
The sixth subindex relates to securities markets and
covers policies to develop domestic bond and
equity markets, including (i) the creation of basic
frameworks such as the auctioning of T-bills,
or the establishment of a security commission;
(ii) policies to further establish securities markets
such as tax exemptions, introduction of medium
and long-term government bonds to establish a
benchmark for the yield curve, or the introduction
of a primary dealer system; (iii) policies to
develop derivative markets or to create an
institutional investor’s base; and (d) policies to
permit access to the domestic stock market by
nonresidents. The subindices are aggregated with
equal weights. Each subindex is coded from zero
(fully repressed) to three (fully liberalized).
 External Capital Qualitative indicators of restrictions on financial Abiad and others (2008), 1973 2005 72 91
Account Liberalization: credits and personal capital transactions of following the methodology in
Aggregate residents and financial credits to nonresidents, as Abiad and Mody (2005), based
well as the use of multiple exchange rates. Index on various IMF reports and
coded from zero (fully repressed) to three working papers, central bank
(fully liberalized). websites, and others.
External Capital Indicators measuring the intensity of legal restrictions Based on the methodology in 1960 2005 50 65
Account Liberalization: on residents', respectively nonresidents', ability to Quinn (1997) and Quinn and
Residents vs. move capital into and out of a country. Index Toyoda (2007), drawing on
Nonresidents originally coded from zero (fully repressed) to 50 information contained in the
(fully liberalized). Fund's AREAER.
Real Sector
Trade Liberalization
Tariff Rates Average tariff rates, with missing values extrapolated Various sources, including IMF, 1960 2005 47 142
using implicit weighted tariff rates. Index World Bank, WTO, UN, and
normalized to be between zero and unity: zero the academic literature
means the tariff rates are 60 percent or higher, (particularly Clemens and
while unity means the tariff rates are zero. Willamson, 2004).
Current-Account An indicator of how compliant a government is with Based on the methodology in 1960 2005 50 65
Restrictions its obligations under the IMF’s Article VIII to free Quinn (1997) and Quinn and
from government restriction the proceeds from Toyoda (2007) drawing on
international trade in goods and services. The information contained in the
index represents the sum of two subcomponents, Fund’s AREAER.
dealing with restrictions on trade in visibles, as well
as in invisibles (financial and other services). It
distinguishes between restrictions on residents
(receipts for exports) and on nonresidents
(payments for imports). Although the index
measures restrictions on the proceeds from
transactions, rather than on the underlying
transactions, many countries in practice use
restrictions on trade proceeds as a type of trade
restriction. The index is scored between zero
and 8 in half-integer units, with 8 indicating
full compliance.
(continued next page)
45
46

Table 1.C2 (continued)

Coverage
MIN # of MAX # of
countries in countries in
Reform Indices Description Source Start Year End Year any year any year
Product Markets
Telecom and Electricity Simple average of the electricity and telecom Based on various existing studies 1960 2003 106 108
Industries markets subindices, which are constructed, and datasets as well as national
in turn, from scores along three dimensions. legislation and other official
For electricity, they capture: (i) the degree of documents.
unbundling of generation, transmission, and
distribution; (ii) whether a regulator other than
government has been established; and (iii) whether
the wholesale market has been liberalized. For
telecom, they capture: (i) the degree of
competition in local services; (ii) whether a
regulator other than government has been
established; and (iii) the degree of liberalization
of interconnection changes. Indices are coded
with values ranging from zero (not liberalized)
to two (completely liberalized).
Agriculture The index captures intervention in the market for Based on IMF commodities data, 1960 2003 96 104
the main agricultural export commodity in each various existing studies and
country. As data limitations preclude coding datasets, and national legislation
separate dimensions of intervention, the index and other official documents.
provides a summary measure. Each country-year
pair can take four values: (i) zero (public monopoly
or monopsony in production, transportation,
or marketing, e.g., export marketing boards);
(ii) one-third (administered prices); (iii) two-thirds
(public ownership of relevant producers or
concession requirements); and (iv) one (no public
intervention).

Source: Prati, Onorato, and Papageorgiou (2010).

ANNEX 1.D NEW DATASET ON STRUCTURAL REFORMS

A discussion of structural reforms, their determinants, and
their economic impact requires solid data on structural
policies and how they have changed (IMF 2009). A new
dataset was compiled by the Research Department of the
IMF that brings together information on a variety of struc-
tural reforms in both the real and financial sectors of the
economy. The dataset spans about 30 years (1973–2005)
and 90 countries, selected on the basis of data availability.
The key advantage of these new indexes over those used in
previous work, such as IMF (2004), is that they cover both a
long sample period and a broad range of countries, includ-
ing advanced and developing economies.
Among the indicators of real sector reforms, two separate
indexes measure trade openness. The first covers average
tariff rates; the second covers restrictions on other current
account transactions, including payments and receipts on
exports and imports of goods and services.9 Two separate
indexes measure product market liberalization. The first
covers the network industries, specifically the degree of lib-
eralization of the electricity and telecom markets. The sec-
ond covers the agriculture sector. It aims to capture public
intervention in the market for the main agricultural export
commodity in each country.
Among the indicators of financial sector reforms, the
index of capital account openness measures restrictions on
financial transactions of residents and nonresidents, as well
as the use of multiple exchange rates. The index of domes-
tic financial liberalization is an average of six subindexes,
covering credit controls, such as directed credit; interest rate
controls, such as floors or ceilings; entry barriers in the
banking sector, such as licensing requirements or limits on
foreign banks; competition restrictions, such as limits on
branches; the degree of state ownership; and aggregate
credit ceilings. All indicators were rescaled to range between
zero and one, with higher values representing greater liber-
alization. Differences in the values of each index across
countries and over time provide useful information on the
variation in the absolute degree of economic liberalization
within each sector. Instead, differences in the value of the
indexes across sectors are not a precise quantitative measure
of whether one sector is more liberalized than another
because of the different methodology used to construct each
index. For instance, a positive difference between the trade
index and the financial index does not necessarily mean that
trade is more liberalized than the financial sector.
As illustrated in figure 1.D1, all indexes trend upward
over time toward a high degree of liberalization. At a sectoral
level, the global liberalization of international trade, capital
movements, and the domestic financial sector has been
fairly steady and gradual over the past three decades,

Figure 1.D1 Indexes of Structural Reform, 1973–2003

(All countries)
0.9

0.8

0.7

0.6

0.5
Indices

0.4

0.3

0.2

0.1

0
19 3
19 4
19 5
19 6
19 7
19 8
79

19 0
19 1
82

19 3
19 4
19 5
19 6
19 7
19 8
19 9
90

19 1
19 2
19 3
19 4
19 5
19 6
19 7
19 8
20 9
20 0
20 1
02

20 3
04
05
7
7
7
7
7
7

8
8

8
8
8
8
8
8
8

9
9
9
9
9
9
9
9
9
0
0

0
19

19

19

19

20

20

Year
Agriculture Capital flows Domestic finance
Networks Trade CA restrictions

Source: IMF 2009.

Note: Data are based on sample of 90 countries for which data were available.

CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 47
 whereas product market liberalization started only about
1990. There have been no global setbacks in the average
degree of liberalization in any sector. Liberalization mea-
sures display significant differences across regions, but all
sectoral indexes have converged across income levels, point-
ing to imitation and “catch-up” effects.
NOTES
1. This chapter draws heavily on Nord et al. (2009). The
authors are grateful to Daehaeng Kim, Saul Lizondo, Roger
Nord, Laure Redifer, and World Bank reviewers for helpful
comments.
2. To the authors’ knowledge, this is the first time Tanzania
has been formally shown to have had an “up-break” in 1996.
Previous studies (Hausmann, Pritchett, and Rodrik 2005;
Berg, Ostry, and Zettelmeyer 2008) failed to identify a break,
probably because they used shorter time series.
3. Excellent descriptions of the reforms undertaken during
these periods from varying perspectives can be found in
Maliyamkono and Mason (2006), Mtei (2010), Mwase and
Ndulu (2008), and Nord et al. (2009).
4. A World Bank (2007) study identifies increases in higher
education attainment as an important contributor to
growth.
5. The tax system has not been fully predictable for investors,
because of fairly frequent changes in administration—often
designed to close loopholes—and the existence of a large
number of tax exemptions granted to particular sectors,
subcategories of investors, and specific enterprises.
6. Government programs during this period, supported by
the IMF (first under the Poverty Reduction and Growth
Facility, subsequently under the Policy Support Instrument),
were designed in such a way that a shortfall in foreign pro-
gram financing automatically triggered the possibility of
resort to high domestic financing subject to a pre-agreed
upper limit; in the event that foreign program financing
exceeded targets, the government was unconstrained in its
ability to spend the additional financing received.
7. The bank closures were Meridien Bank (in 1995),
Tanzania Housing Bank (in 1995), Greenland Bank (in
1999), and Delphis Bank (in 2003). Depositors were fully
compensated in all cases, with the resolutions of both
Meridien Bank and Delphis Bank facilitated by the trans-
fers of deposits to other banks operating in Tanzania.
8. The models predict tax potential through a regression of
the tax to GDP ratio on observable structural characteristics
of the economy, such as the size of the agricultural sector.
9. This index—measuring how compliant a country is
with its obligations under the IMF’s Article VIII to free from
restrictions the proceeds from international trade in goods
48 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
and services—captures only some nontariff barriers to
trade. For other nontariff barriers, such as quotas and sub-
sidies, there is not a sufficiently long time series to be used
in the analysis.
REFERENCES
Abbas, S., M. Ali, and Y. Sobolev. 2008. “High and Volatile
Treasury Yields in Tanzania: The Role of Strategic Bid-
ding and Auction Microstructure.” IMF Working Paper
08/81, International Monetary Fund, Washington, DC.
Bai, J., and P. Perron. 1998. “Estimating and Testing Linear
Models with Multiple Structural Changes.” Econometrica
66: 47–78.
Berg, A., D., J. Ostry, and J. Zettelmeyer. 2008. “What Makes
Growth Sustained?” IMF Working Paper 08/59, Interna-
tional Monetary Fund, Washington, DC.
Drummond, Paulo, A. Mrema, S. Roudet, and M. Saito.
2009. Foreign Exchange Reserve Adequacy in East African
Community Countries. African Departmental Paper
09/01, International Monetary Fund, Washington, DC.
Easterly, W., M. Kremer, L. Pritchett, and L. H. Summers.
1993. “Good Policy or Good Luck? Country Growth Per-
formance and Temporary Shocks.” Journal of Monetary
Economics 32 (3): 459–483.
Gratwick, K. N., and A. Eberhard. 2008. “An Analysis of
Independent Power Projects in Africa: Understanding
Development and Investment Outcomes.” Development
Policy Review 26 (3): 309–38.
Hausmann, R., L. Pritchett, and D. Rodrik. 2005. “Growth
Accelerations.” Journal of Economic Growth 10 (4): 303–29.
Hobdari, Niko. 2008. “Tanzania’s Equilibrium Real Exchange
Rate.” IMF Working Paper 08/138, International Mone-
tary Fund, Washington, DC.
IMF (International Monetary Fund). 2004. “Advancing
Structural Reforms.” World Economic Outlook (April),
Washington, DC.
———. 2009. Structural Reforms and Economic Performance
in Advanced and Developing Countries. Occasional Paper
268, Washington, DC.
———. 2010a. “Sub-Saharan Africa: Back to High Growth?”
Regional Economic Outlook, April, Washington, DC.
———. 2010b. Tanzania: Financial System Stability Assess-
ment Update. IMF Country Report 10/177.
———. 2010c. Tanzania: Seventh Review under the Policy
Support Instrument, Second Review under the Exogenous
Shocks Facility, and Request for a New Three-Year Policy
Support Instrument. IMF Country Report 10/173,
Washington, DC.
———. 2010d. World Economic Outlook, April. Washington,
DC.
Lawson, A., D. Booth, M. Msuya, S. Wangwe, and
T. Williamson. 2005. Does General Budget Support Work?
Evidence from Tanzania. Overseas Development Institute,
London.
Maliyamkono, T. L., and H. Mason. 2006. The Promise. Dar
es Salaam: TEMA Publishers.
Mbelle, A. V. Y., L. Rutasitra, B. Makenda, R. Lokina,
A. Naho, and Y. Aikaeli. 2010. “Analytical Study on the
Drivers of Growth and Implications for Growth Strategy
for Mainland Tanzania,” University of Dar es Salaam,
Dar es Salaam.
Mtei, Edwin. 2009. From Goatherd to Governor. Dar es
Salaam: Mkuki na Nyota Publishers.
Mwase, Nkunde, and Benno Ndulu. 2008. “Tanzania:
Explaining Four Decades of Episodic Growth.” In The
Political Economy of Economic growth in Africa
1960–2000, vol 2., Country Case Studies, ed. Ndulu, B,
S. O’Connell, J. P. Azam, R. Bates, A. Fosu, J. Gunning,
and D. Njinkeu. Cambridge: Cambridge University Press.
National Audit Office. 2010. A Performance Audit Report on
Health Care Centres Efficiency in Tanzania. www.nao.go.tz
Nord, Roger, Yuri Sobolev, David Dunn, Alejandro Hajden-
berg, Niko Hobdari, Samar Maziad, and Stephane
Roudet. 2009. Tanzania: The Story of an African Transi-
tion. International Monetary Fund, Washington, DC.
Ostry, J. D., A. Prati, and A. Spilimbergo. 2009. Structural
Reforms and Economic Performance in Advanced and
Developing Countries. IMF Occasional Paper 268,
Washington, DC.
Prati, Alessandro, Massimiliano Onorato, and Chris Papa-
georgiou. 2010. “Which Reforms Work and under What
Institutional Environment? Evidence from a New
Dataset on Structural Reforms.” International Monetary
Fund, Washington, DC.
CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY
Pritchett, Lant. 2000. “Understanding Patterns of Economic
Growth: Searching for Hills among Plateaus, Mountains,
and Plains.” World Bank Economic Review 14 (2): 221–50.
Redifer, Laure. 2010. “New Financing Sources for Africa’s
Infrastructure Deficit.” IMF Survey July 21, Washington,
DC.
Selassie, Abebe Aemro. 2008. “Beyond Macroeconomic
Stability: The Quest for Industrialization in Uganda.”
IMF Working Paper 08/231, International Monetary
Fund, Washington, DC.
Ter-Minassian, Teresa, Richard Hughes, and Alejandro
Hajdenberg. 2008. “Creating Sustainable Fiscal Space for
Infrastructure: The Case of Tanzania.” IMF Working Paper
08/256, International Monetary Fund, Washington, DC.
United Republic of Tanzania. 2009a. “Millennium Develop-
ment Goals Report, Mid-Way Evaluation: 2000–2008.”
www.tzdpg.or.tz/external/mkukuta-mkuza-review/
studies-tors-cvs.html.
———. 2009b. MKUKUTA Annual Implementation Report
2008/2009, Success in the Midst of Turbulence. Dar es
Salaam.
———. 2010. Public Expenditure Tracking Survey for Pri-
mary and Secondary Education in Mainland Tanzania.
Dar es Salaam.
World Bank. Various years. World Development Indicators.
Washington, DC: World Bank.
———. 2005. Agriculture and Achieving the Millennium
Development Goals. Report 32729-GLB, Agriculture and
Rural Development Department, Washington, DC.
———. 2007. Tanzania: Sustaining and Sharing Economic
Growth. Country Economic Memorandum, Washington,
DC.
———. 2009. Doing Business 2010, Tanzania. Washington,
DC: World Bank.
49
 CHAPTER 2
Building on Growth in Uganda
Sarah Ssewanyana, John Mary Matovu, and Evarist Twimukye
ver the past two decades, Uganda has seen a
O
remarkable turnaround in economic perfor-
mance, with growth averaging about 7.7 percent
a year over the 1997–2007 period. Equally impressive has
been the sharp decline in poverty rates, which fell about
15 percentage points over this period. Improved macroeco-
nomic management and economic reforms contributed to
the country’s strong growth performance.
Although growth and poverty reduction have been
impressive, Uganda has experienced worsening income dis-
tribution; a decline in the relative importance of agriculture
to overall gross domestic product (GDP); a growing youth
population, with increasing unemployment; and a low rate
of urbanization. There has also been limited structural
transformation of the economy, a reflection that growth has
come largely from the services sector, which employs the
highly skilled, rather than from the agricultural sector,
which still employs 70 percent of the population. Notwith-
standing the considerable progress in diversifying its export
base away from coffee, Uganda still remains a primary
commodity exporter, with limited value addition to its
major exports.
GROWTH EXPERIENCE
Uganda achieved impressive economic growth over the past
two decades, with positive per capita GDP growth since
1987 and stronger growth than the continent as a whole
(figure 2.1 and table 2.1). Despite this improvement,
Uganda’s per capita GDP at purchasing power parity
remains about half that of Sub-Saharan African as a whole.
During the decade following the end of political instabil-
ity and civil war in 1986, Uganda’s economy grew at an aver-
age rate of 7.7 percent a year. (Economic growth declined by
1.4 percentage points between 2008/09 and 2009/10.) Initially,
economic growth was driven by postwar recovery and recon-
struction. Since the early 1990s it has been driven by compre-
hensive macroeconomic and structural reforms. Investment
growth also remained strong, with private investment rising
by an estimated 17 percent and public investment rising
15 percent (table 2.2). Private investment growth was led by
construction (AfDB 2009).
Economic growth has been export led, with the share
of exports in GDP rising over the past two decades. The
expanding regional market for Uganda’s food and manu-
factured products has boosted exports during the past five
years, a reflection of the dividend enjoyed by Uganda’s
neighbors, whose demand for Ugandan goods has
increased.
Growth has been driven by the services sector, which has
accounted for almost half of GDP since 2001/02 (table 2.3).
Before the global financial crisis, the key subsectors driving
growth in the services sector included financial services,
transport and communications, public administration, and
defense (see annex table 2.A1). The agricultural sector grew
more slowly than the other sectors between 2001/02 and
2009/10, accounting for about 24 percent of annual GDP on
average over the period (table 2.3).
51
 Figure 2.1 Annual GDP Growth in Uganda, 1983–2007

12

10

Per capita GDP (dollars)

8
1,000
6

4 500
Percent

0 0

–2

–4

–6

–8
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
00
01
02
03
04
05
06
07
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
20
20
20
20
20
20
20
20
GDP per capita, (purchasing power parity in constant 2005 dollars) GDP growth (annual percent)
GDP per capita growth (annual percent)

Source: World Bank 2009.

Table 2.1 Annual GDP Growth in Selected Countries in Sub-Saharan Africa, 1997–2009

Sub-Saharan
Year Ghana Kenya Mauritius Mozambique South Africa Tanzania Tunisia Uganda Africa
1997 4.20 0.47 5.69 10.24 2.65 3.53 5.44 5.10 3.58
(1.59) (–2.23) (4.37) (7.26) (0.32) (0.84) (4.00) (1.94) (0.87)
1998 4.70 3.29 6.07 10.78 0.52 3.71 4.78 4.91 2.42
(2.15) (0.59) (4.96) (7.96) (–1.82) (1.12) (3.46) (1.77) (–0.24)
1999 4.40 2.31 2.61 8.12 2.36 3.53 6.05 8.05 2.46
(1.90) (–0.33) (1.31) (5.39) (–0.08) (0.97) (4.68) (11.11) (–0.19)
2000 3.70 0.60 9.03 1.09 4.15 5.10 4.70 3.14 3.53
(1.24) (–1.98) (7.96) (–1.52) (1.61) (2.46) (3.52) (–0.01) (0.87)
2001 4.00 3.78 2.57 11.9 2.74 6.24 4.92 5.18 3.62
(1.56) (1.11) (1.46) (8.93) (0.65) (3.53) (3.73) (1.92) (1.00)
2002 4.50 0.55 2.11 8.82 3.67 7.24 1.65 8.73 3.45
(2.07) (–2.03) (1.24) (5.92) (2.25) (4.46) (0.53) (5.31) (0.87)
2003 5.20 2.93 3.66 6.02 2.95 5.67 5.56 6.47 4.23
(2.79) (0.29) (2.59) (3.21) (1.65) (2.84) (4.94) (3.09) (1.67)
2004 5.60 5.10 5.75 7.88 4.55 6.73 6.04 6.81 6.19
(3.23) (2.40) (4.84) (5.08) (3.33) (3.88) (5.05) (3.39) (3.60)
2005 5.90 5.91 1.24 8.39 5.28 7.37 3.98 6.33 5.72
(3.58) (3.17) (0.44) (5.66) (4.09) (4.46) (2.98) (2.92) (3.15)
2006 6.40 6.32 3.95 8.68 5.60 6.74 5.66 10.78 6.17
(4.13) (3.56) (3.14) (6.04) (4.43) (3.79) (4.63) (7.23) (3.59)
2007 6.46 7.01 5.52 7.28 5.49 7.15 6.33 8.41 6.60
(4.24) (4.22) (4.87) (4.75) (4.34) (4.15) (5.32) (4.92) (4.02)
2008 8.43 1.55 5.09 6.74 3.68 7.44 4.64 8.71 5.26
(6.20) (–1.09) (4.41) (4.29) (2.54) (4.39) (3.60) (5.21) (2.72)
2009 4.66 2.59 2.14 6.33 –1.78 5.50 3.13 7.06 1.65
(2.52) (–0.08) (1.62) (3.96) (–2.83) (2.48) (2.10) (3.62) (–0.80)

Source: World Bank 2010.

Note: Figures in parentheses are per capita growth rates.

52 CHAPTER 2: BUILDING ON GROWTH IN UGANDA

 Table 2.2 Components of GDP Growth in Uganda, 2001–09
(percent, in 2002 prices)
Component 2001 2002 2003 2004 2005 2006 2007 2008 2009
Consumption
Private 6.3 6.1 2.8 3.4 8.1 12.3 2.9 8.2 11.5
Public 2.9 5.2 5.2 3.7 4.5 3.7 –1.5 1.7 –1.9
Investment
Private 10 14.1 16.1 14.6 23.5 10.3 14.6 12.0 1.1
Public 0.3 –8.1 2.3 11.6 9.3 13.5 19.6 –15.3 28.8
Trade
Exports 17.2 7 6.6 20.9 21.5 –6.3 53.8 45.0 –12.0
Imports 12.8 7.5 6.2 9.6 16.8 17.2 15.7 17.6 4.6

Source: Uganda Bureau of Statistics 2010.

Table 2.3 Sectoral Contributions to GDP and Growth Rates in Uganda, 2001/02–2008/09
(percent, in 2002 prices)
Item 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10
Sectoral contributions to GDP (in current prices)
Agriculture 23.1 22.1 21.1 20.2 24.1 22.3 21.4 23.1 23.9
Industry 22.0 22.6 22.8 24.0 22.8 25.1 25.8 24.7 24.6
Services 48.3 48.6 49.1 49.0 47.2 47.0 46.9 46.4 45.4
Growth rate
Total GDP 8.5 6.5 6.8 6.3 10.8 8.4 8.7 7.2 5.8
Per capita GDP 5.1 3.1 3.4 3.0 7.3 5.0 5.3 3.8 2.4
Agriculture 7.1 2.1 1.6 2.0 0.5 0.1 1.3 2.5 2.1
Industry 7.4 9.5 8 11.6 14.7 9.6 8.8 5.8 8.9
Services 11.0 7.4 7.9 6.2 12.2 8.0 9.7 8.8 5.8

Source: Uganda Bureau of Statistics 2010.

The contribution of industry to GDP—which ranged
from 22.0 to 24.6 percent over the period—is well below the
35 percent benchmark for countries graduating from low- to
middle-income status (Bevan and others 2003). Growth in the
sector fell from 9.1 percent in 2007/08 to 5.8 percent in
2008/09, largely as a result of the global economic crisis, which
reduced the remittances that had fueled a construction boom
in Uganda. The decline caused the share of the construction
subsector in GDP to fall from 10.8 percent in 2007/08 to
3.7 percent in 2008/09 (see annex table 2.A1). The slowdown
was also a result of the increase in the costs of imported
inputs arising from the depreciation of the Ugandan shilling.
The contribution of the manufacturing subsector to overall
GDP was well below that of the construction subsector.
Structural transformation of the economy
Uganda’s economy has been growing rapidly, but growth
has failed to create enough jobs for the ever-increasing labor
force. Evidence from the Uganda National Household Sur-
veys shows that as a result of impressive economic growth
during the 1990s, the population experienced important
broad-based welfare gains in terms of consumption
increases. The gains were not equally distributed across
social groups and spatially. Overall, the pattern of growth
has been skewed, with growth taking place in subsectors
such as telecommunications and finance, which employ
highly skilled people, not the poor.
The mismatch between the contribution of the different
sectors to GDP and the proportion of the population that
derives its livelihood from the sectors has serious implica-
tions for the level of unemployment and underemployment
in Uganda. Although the contribution of agriculture to
overall GDP has remained lower than that of services or
industry, its share in employment remains high. Services,
which account for almost half of GDP, employ only about
24 percent of the population. Employment in industry,
which accounts for more than a quarter of GDP, is also very
minimal, at about 8 percent of the population (table 2.4).
Agricultural performance
Because more than two-thirds of Ugandans work in agricul-
ture, development of the Ugandan economy is closely linked

CHAPTER 2: BUILDING ON GROWTH IN UGANDA 53

 to transformation of the agricultural sector. Agriculture in
Uganda is still characterized by low productivity, mainly as
a result of poor inputs, undeveloped value chains, and low
public and private investment in the sector. The lack of sus-
tained agricultural growth and the slow process of diversi-
fication in agriculture pose serious threats to poverty
reduction efforts.
The reforms of the early 1990s, especially the dismantling
of the agricultural public enterprises and liberalization of the
economy, led to greater participation of the private sector in
marketing agricultural produce. The agricultural reforms
implemented since the mid-1990s, however, have largely
benefited only a small fraction of farmers, particularly richer
and better-educated farmers, who have been able to diversify
their agricultural production. Once these efficiency gains
were exploited, other innovations were needed to maintain
growth in the sector (Okidi and others 2007).
Only 50 percent of agricultural production in Uganda is
sold on markets. With the exception of coffee farmers and
farmers engaged in other tradables or niche markets
(vanilla, fruit, tomatoes), most smallholders in Uganda are
still engaged in subsistence farming. For the most part,
Ugandan farmers remain poor and out of mainstream eco-
nomic activity.
then, Uganda has diversified its exports base to include
larger shares of flowers, fishing, and other agricultural
exports (see annex table 2.A2). Revenue from noncoffee
exports increased by more than sixfold between 1997/98
and 2008/09, rising from $189.6 million to $1,199.6 million
(figure 2.3).
Trade liberalization was designed to reverse and even
eliminate the trade deficit by increasing export earnings
and curtailing the demand for imports. Incentives for
export-oriented trade and market-determined exchange
rate policies were expected to encourage both traditional
and nontraditional exports. Nevertheless, merchandise
exports continued to decline throughout the liberalization
period (1987–92), partly because the manufacturing sector
had shrunk as a result of the economic mismanagement of
Figure 2.2 Exports as Percentage of GDP in Uganda,
1998/99–2008/09
25
20
15
Percent

10
Export performance
5
Exports as a share of GDP have increased over time in
Uganda (see figure 2.2). Before the liberalization of the 0
economy and the emphasis on import substitution and
9

20 5
6

20 7

20 8
9
/9

/0

/0

/0

/0

/0

/0

/0

/0

/0

/0
98

99

00

01

02

03

04

05

06

07

08
export diversification in the 1990s, Uganda depended
19

19

20

20

20

20

20

20
mainly on coffee as its main export. This dependence on a Source: Bank of Uganda 2009.
single commodity was a major constraint to terms-of-trade
growth, especially when world coffee prices dropped, as they
did in the mid-1990s.
Figure 2.3 Coffee, Noncoffee, and Total Exports from
To insulate the economy from adverse terms of trade
Uganda, 1997/98–2008/09
and instability in export earnings associated with com-
modity concentration, the government adopted a policy 3,500
shift in 1987 that sought to diversify the exports base to 3,000
Millions of dollars

include nontraditional (mainly agricultural) exports. Since 2,500

2,000
1,500
Table 2.4 Distribution of Employment in Uganda, 1,000
by Sector and Gender, 2003 500
(percent) 0
8

1
2

9
/9

/9

/0

/0
/0

/0

/0

/0

/0

/0

/0

/0

Sector Women Men Total

97

98

99

00
01

02

03

04

05

06

07

08

Agriculture 76 62 69 Total exports Coffee

Industry 5 10 8
Noncoffee Informal cross-border
Services 19 28 24
Source: World Bank 2009. Source: Bank of Uganda 2008/09.

54 CHAPTER 2: BUILDING ON GROWTH IN UGANDA

earlier regimes. Since then, the value of exports has Figure 2.5 Trade Deficit as Percentage of GDP in
improved markedly. Uganda, 1998/99–2008/09
Alongside the increase in noncoffee exports has been a
huge increase in the importance of informal cross-border 10
9
trade, which rose from 1.3 percent of total exports in
8
2002/03 to 50.2 percent by 2008/09—a significant develop- 7
ment in the wake of the global financial crisis, which was 6

Percent
beginning to affect traditional exports, especially coffee. 5
Including informal cross-border trade, the share of indus- 4
3
trial products in total exports increased from 43.8 percent in 2
2007/08 to 54.9 percent in 2008/09 (figure 2.4). As new 1
emerging export markets within the region stabilize, it will 0
be difficult for Uganda to sustain the recent rate of indus-

19 9

20 0

20 1

20 2

20 3

20 4

20 5

20 6

20 7

20 8
9
/9

/0

/0

/0

/0

/0

/0

/0

/0

/0

/0
98

99

00

01

02

03

04

05

06

07

08
trial product export growth.

Most primary commodities earn a fraction of what
they would earn if they had been processed. Conse-
quently, although the shilling depreciated against the U.S
dollar in real terms (from U Sh 558 per dollar in 1987 to
U Sh 2,000 in 2008), implying higher domestic producer
prices in Uganda shillings, the trade deficit ballooned,
from $446.7 million in 1997/98 to $936.3 million in
2008/09. The increase in the trade deficit reflected the low
returns from exports caused by the deteriorating terms of
trade and the large increase in the value of imports caused
by the depreciation of the shilling. The trade deficit as a
share of GDP has declined since 2005/06 (figure 2.5).
Uganda will potentially benefit from the expanded East
African Common Market, estimated at 120 million people,
which allows unimpeded movement of labor, capital, and
other services across borders within Burundi, Kenya,
Rwanda, Tanzania, and Uganda. However, trade with the
largest economies in the union (Kenya and Tanzania) has
been one sided, with limited exports from Uganda to the
two countries. Uganda’s trade has been boosted mainly by
Figure 2.4 Industrial Products as Percentage of Total
Exports in Uganda, 1997/98–2008/09
60
50
40
Percent
19
Source: Bank of Uganda 2009.
trading with postconflict economies in the region, includ-
ing Southern Sudan, the Democratic Republic of Congo,
and Rwanda. As these countries stabilize, the trade deficit
of Uganda with its major East African Community partners
will have to be addressed by increasing exports to Kenya
and Tanzania.
Impact of growth on poverty reduction and
other millennium development goals
Uganda is one of the few Sub-Saharan African countries
to achieve the first MDG of halving extreme poverty
before 2015 (annex table 2.A3). The proportion of the
population living below the absolute poverty line declined
from 56.4 percent in 1992/93 to 31.1 percent in 2005/06
and to 24.5 percent in 2009/10 (table 2.5).1 Income
poverty remains a key development challenge, however,
with the absolute number of poor people declining only
marginally, from 9.8 million in 1992 to 8.4 million in
2005/06 and to 7.5 million in 2009/10, as a result of a pop-
ulation growth rate of 3.2 percent a year. The majority of
the poor live in rural areas, in particular in northern
Uganda (Ssewanyana and Okidi 2007; Uganda Bureau of
Statistics 2010).
The regional ranking of other poverty measures in

30 Uganda is identical to that of the headcount index, although

20 there are growing differences in the poverty gap between
10 northern Uganda and the rest of the country. Overall, the
0 poverty gap dropped faster than the poverty headcount,
implying rising mean consumption by Uganda’s poor. The
8

9
/9

/9

/0

/0

/0

/0

/0

/0

/0

/0

/0

/0
97

98

99

00

01

02

03

04

05

06

07

08

recovery in the agricultural sector, especially the food

All exports Industrial formal Industrial informal
crop subsector; return to peace in northern Uganda and
Source: Bank of Uganda 2008/09. some parts of the eastern region; and resettlement of the

CHAPTER 2: BUILDING ON GROWTH IN UGANDA 55

 Table 2.5 Poverty Headcount and Income Inequality Estimates for Uganda, 1992/93–2005/06
Poverty headcount Gini coefficient
Region 1992/93 2002/03 2005/06 2009/10 1992/93 2002/03 2005/06 2009/10
National 56.4 38.8 31.1 24.5 0.365 0.428 0.408 0.426
Rural 60.3 42.7 34.2 27.2 0.328 0.363 0.363 0.375
Urban 28.8 14.4 13.7 9.1 0.396 0.483 0.432 0.447
Central 45.6 22.3 16.4 10.7 0.395 0.460 0.417 0.451
East 58.8 46.0 35.9 24.3 0.327 0.365 0.354 0.319
North 73.5 63.0 60.7 46.2 0.345 0.350 0.331 0.367
West 52.7 32.9 20.5 21.8 0.319 0.359 0.342 0.375

Source: Ssewanyana and Okidi 2007; Uganda Bureau of Statistics 2010.

Note: The poverty headcount is the percentage of people estimated to be living in households with real private consumption per adult equivalent
below the poverty line for their region.

internally displaced persons partly accounted for the
decline in poverty between 2005/06 and 2009/10.
Although absolute income poverty fell during the past
two decades, the distribution of income worsened. Income
inequality as measured by the Gini coefficient increased
from 0.365 in 1992/93 to 0.408 in 2005/06. Growing income
disparities are evident between 2005/06 and 2009/10, when
the Gini coefficient reached 0.426. Growth rates for con-
sumption grew more rapidly among the richer quintiles,
increasing inequality (Ssewanyana and Okidi 2007; Uganda
Bureau of Statistics 2010). Income inequality between rural
and urban areas and between regions widened between
2005/06 and 2009/10.
Increasing inequality has slowed the rate of poverty
reduction in Uganda. For every 1 percent decrease in growth,
the percentage of people living below the poverty line will
increase by 2 percent, holding income distribution constant
(Ssewanyana 2009). This means that the economy has to
grow by at least 7 percent and household consumption has
to rise at least 4 percent if Uganda is to avoid reversals in its
poverty reduction efforts.
Uganda is likely to attain MDG 3 (promoting gender
equality and empowering women), MDG 6 (combating
HIV/AIDS), MDG 7 (ensuring environmental sustainability),
and MDG 8 (developing a global partnership for develop-
ment). The stagnation in net primary school enrollment since
2003 at about 85 percent is a clear indication that intensified
efforts are required if Uganda is to meet MDG 2, however,
and attainment of MDG 4 (reducing child mortality) and
MDG 5 (improving maternal health) is unlikely even with
improved policies, institutions, and funding (UNDP 2010).
REFORMS DRIVING GROWTH
The past two decades have seen tremendous economic
transformation in Uganda, fueled mainly by good policies
involving careful sequencing and determined implementa-
tion. The World Bank has referred to Uganda’s efforts as “the
most far-reaching stabilization and structural reform pro-
gram in Africa, and one of the most comprehensive reform
efforts in the world” (World Bank 2007, 4).
The first reform was the Economic Recovery Program,
introduced in 1987, with support from the World Bank and
International Monetary Fund (IMF), which focused on
price stabilization and liberalization. Policies under this
program included currency reform, devaluation, liberaliza-
tion of domestic prices, and conversion to a floating
exchange rate regime (in 1993).
The next set of reforms involved the adoption of the
structural adjustment program that was meant to free up
markets and create price incentives, stimulate private
investment, and encourage competition. Reforms under
this program included the abolition of marketing boards,
the privatization or abolition of parastatals, and the
establishment of the Uganda Investment Authority. This
period was characterized by sustained macroeconomic
stabilization, adjustment, and structural reform efforts
that affected almost all sectors of the economy. Policies
mainly involved the macroeconomic stabilization process,
price liberalization, financial sector liberalization, public
enterprises reform, and civil service reform. In addition to
changing and stabilizing the structure of the economy,
reorientation of pricing and marketing policies, restarting
of economic growth, and strengthening the institutional
framework constituted the major cornerstones of the
program. To achieve these ends, the Economic Recovery
Program focused on ensuring macroeconomic stability;
liberalizing the foreign exchange system, trade, prices, and
marketing systems; improving the incentive structure
and business climate to promote savings mobilization and
investment; and rehabilitating the economic, social, and
institutional infrastructure.

56 CHAPTER 2: BUILDING ON GROWTH IN UGANDA

 With the economy back on its footing, in 1997 the govern-
ment introduced the Poverty Eradication Action Plan (PEAP),
a multisectoral program aimed at reducing poverty. Policies in
this program included the Plan for the Modernization of
Agriculture (PMA), which sought to address agricultural
constraints to production and to turn agriculture commer-
cial. The plan has not been as successful as envisaged, mainly
because it was too broad and in some cases ambiguous, with
several programs having too many pillars when the focus
should have been on enhancing agriculture productivity.
The National Agricultural Advisory Services, the flagship
agricultural productivity enhancement program within
PMA, was beset by inefficiencies and other implementation
problems that limited its impact on agriculture.2
Other sectoral reforms that contributed to the liberaliza-
tion and stabilization of the economy included the
Medium-Term Competitive Strategy for the Private Sector,
the Strategic Export Program, and the Strategic Export
Intervention Program. These policies were accompanied by
important institutional reforms, such as decentralization
efforts, the abolition of state-owned marketing boards, and
the restructuring of the public administration.
The 1990s saw a substantial reversal in the decline of the
economy that had characterized the 1970s and early 1980s,
suggesting that reform worked. Confidence in the economy
was restored, spurring substantial inflows of aid and foreign
direct investment and a reversal of capital flight (figure 2.6).
Most economic indicators rebounded, and by 1996 the
economy had recovered to its nominal 1971 dollar per
capita GDP (World Bank 2007).
To sustain rapid economic growth, the government
needed to reorient expenditures toward social sectors and
increase spending on infrastructure. Social policy spending
Figure 2.6 Official Development Aid as Percentage of
GDP in Uganda, 1998/99–2008/09
was aimed at spurring growth as well as the level of produc-
tivity of assets of poor people. Policies undertaken under
the framework of the PEAP included universal primary and
secondary education, intended both to raise education indi-
cators and to remove the financial burden of education
from parents as a means of reducing poverty. Efforts were
also made to increase health coverage, by constructing
health centers in all subcounties and parishes.
Examination of the composition of the public budget for
most of the 1990s and 2000s reveals that total government
expenditure steadily increased, from about 18.6 percent of
GDP in 1992/93 to about 32.0 percent in 2008/09. Evidence
suggests that increased funding to social services has bene-
fited the poor, especially in rural areas (Kappel, Lay, and
Steiner 2005).
With reasonable progress in the social sectors, in the late
2000s the government began addressing the infrastructural
constraints that have dogged the country since indepen-
dence. Since 2007/08 substantial resources have been com-
mitted to the rehabilitation and construction of roads and
hydroelectric dams. The government has prioritized the
building of roads in the medium and long term, with more
than 20 trunk roads planned or in the process of being built
in the next 20 years (MFPED 2009). The share of total cen-
tral government budget allocated to works and transport
rose from 10.0 percent in 2005/06 to 18.4 percent in
2009/10, and the share of the budget allocated to energy and
mineral development rose from 3.5 to 10.3 percent. Taken
together, allocations for works, transport, and energy repre-
sented almost a third of the 2009/10 budget—a substantial
figure given Uganda’s historically low investments in infra-
structure. Although issues about absorption capacity and
quality of infrastructure remain, there is a perception that
the government’s tightened focus on improving infrastruc-
ture will help reduce Uganda’s perennial infrastructural
shortages, giving the economy a major boost.

14 CHALLENGES TO DEVELOPMENT
12
Policy makers need to address a variety of challenges in
10
reducing poverty and spurring growth. In addition to the
Percent

8 challenges identified in this section, they need to ensure that

6 growth is inclusive. Failure to achieve equity is likely to
4 exacerbate unemployment and lead to social unrest.
2
0 Slow progress in reducing poverty and slowing
population growth
9

9
/9

/0

/0

/0

/0

/0

/0

/0

/0

/0

/0
98

99

00

01

02

03

04

05

06

07

08
19

19

20

20

20

20

20

20

20

20

20

Although Uganda has already achieved MDG 1 (halving

Source: Bank of Uganda 2009. extreme poverty), poverty remains high, especially in the

CHAPTER 2: BUILDING ON GROWTH IN UGANDA 57

 northern region. Poverty reduction interventions, especially
the Peace and Reconciliation Development Plan (PRDP),
in this part of the country need to be restructured to ensure
that they reach and benefit vulnerable groups.
Even at the average GDP growth rate of 6.9 percent
attained during the 1990s, it would take Uganda about 20
years to double average per capita income. With population
growth rate at 3.2 percent a year, one of the highest popula-
tion growth rates in the world, it is going to be very difficult
to reduce poverty (World Bank 2007).
Uganda’s 3.2 percent annual population growth thwarts
development, especially in social services provision and
social outcomes. Unless efforts at reducing the high popula-
tion growth rates are intensified, many of the MDGs will be
difficult to achieve.
Inadequate transformation of the
agriculture sector
Land reform is a critical factor that will ensure sustainable
agricultural productivity and facilitate commerce and
agriculture. Such reform would involve the government
buying out landlords in order to consolidate the small
plots on which many rural farmers depend to enable com-
mercial agriculture. The government could also try to
draw the rural population to urban centers through more
planned urban development.
Other factors explaining low agricultural productivity
include poor inputs, undeveloped value chains, and low
public and private investment in the sector. The issue of low
productivity in agriculture will need to be addressed by, for
instance, accelerating the creation of on-farm and off-farm
processing zones to add value to agricultural commodities,
financing activities along the entire value chain as opposed to
funding on-farm activities only, and land reforms that will
allow commercial agricultural production as articulated in the
National Development Plan (Republic of Uganda 2010).
Weak infrastructure
Uganda’s infrastructure is among the worst in the world
(Republic of Uganda 2010; World Bank 2007). The National
Development Plan identifies weak infrastructure as one of
the key binding constraints in Uganda. Roads, power and
railways are all below those of Uganda’s neighbors, with
grave implications for the economy.3 The National Develop-
ment Plan identifies the most challenging infrastructural
impediments as power, transport, and access to and the high
cost of finance, in that order. Efforts to address infrastructure
58 CHAPTER 2: BUILDING ON GROWTH IN UGANDA
constraints by increasing sector funding represent a move
in the right direction and should be followed by prudent
use of the resources by responsible agencies to avoid leak-
ages, which have often derailed service delivery. Regional
infrastructure also remains a key challenge, especially for
boosting regional trade.
Lack of sufficient private sector development
Private sector growth has been impressive since liberaliza-
tion, but Uganda’s economy remains dominated by small
firms that usually employ fewer than five people, making
it difficult to absorb the growing number of graduates and
exacerbating the youth unemployment problem. In addi-
tion, although the private sector has played a significant
role in areas such as education and health services, it has
not been prominent in other sectors that require signifi-
cant investments, such as energy, and even in sectors in
which the private sector is active, the impact of its activi-
ties on employment creation has been limited. Enlarging
the role of the private sector calls for more active partici-
pation of the government under private-public partner-
ship arrangements.
Government development programs and successive gen-
erations of Country Strategy Papers have emphasized the
development of the private sector as a major goal for the
country. But progress has been slow in this area, partly
because of weak human and institutional capacity, which
limits private sector participation in execution of contracted
or self-initiated projects. Further limiting the growth of the
private sector is the poor state of the financial sector, which
limits access to credit. Equity finance is also lacking, and
small businesses have limited access to commercial banking
facilities. It is important that the government expedites the
recapitalization of the Uganda Development Corporation
and the Uganda Development Bank to make private sector
access to credit easier.
Narrow export base and terms-of-trade
vulnerability
Despite the diversification of its export base, Uganda
remains heavily dependent on primary commodities. Diver-
sification of the export base is of paramount importance.
The factors that continue to constrain export diversification
include the primary and low-value-added nature of
Uganda’s exports, poor product quality, and poor regulation
standards, which inhibit competition in marketing and
export of primary commodities.
Low tax revenue
Tax revenue in Uganda is low, the result mainly of untaxed
sectors, especially informal businesses and some agricul-
tural activities, and tax evasion (Sennoga, Matovu, and
Twimukye 2009). The lack of revenue has translated into an
unending dependence on foreign aid, which accounted for
about 32 percent of the budget in 2008/09. In 2007/2008 tax
revenue amounted to about 12 percent of GDP. This figure
is low relative to Uganda’s neighbors (tax revenue was about
17 percent of GDP in Tanzania and about 27 percent of
GDP in Kenya for the same period). Lack of adequate tax
revenues constrains government operations and weakens
economic management. It also increases government
reliance on foreign aid to finance development, making eco-
nomic management more difficult.
The recent discovery of oil in Uganda could significantly
boost the resources mobilized domestically, widening
resources available to finance Uganda’s development agenda.
The government could also try to expand the tax base, by
targeting sectors that are currently untaxed, especially the
informal sector, and setting up regulation to reduce tax
evasion.4
Corruption and weak governance
The government has made numerous efforts to strengthen
good governance, by allowing civil society organizations to
participate more in planning and budgeting and by setting
up institutions to strengthen accountability. Such initiatives
included the creation, in 2008, of the anticorruption court
to handle corruption-related cases and the value-for-money
audit unit in the Auditor General’s office to check on gov-
ernment spending at all levels. Uganda has also adopted a
national anticorruption strategy.
These efforts notwithstanding, the perception remains
that corruption is rampant, and the government has come
under pressure from several quarters to show more political
will to address the problem. Uganda’s ranking on the
Ibrahim Index of African governance improved from 27th
in 2007 to 19th in 2008. But in May 2009, Transparency
International ranked Uganda the third most corrupt coun-
try in the world. And although there has been considerable
progress in tax administration with the formation of the
Uganda Revenue Authority in 1991, tax policy and adminis-
tration remain issues as they relate to revenue mobilization,
partly because of corruption within the tax body. In 2010
Transparency International ranked Uganda’s revenue
authority the second most corrupt tax body in East Africa
(Transparency International 2010).
More needs to be done to improve governance in
Uganda, where civil society organizations are weak and the
legislature often fails to rein in Uganda’s powerful executive.
Unless governance is strengthened, even increased resources
may fail to bring meaningful development to Uganda.
Weak human and institutional capability
Uganda’s Five-Year National Development Plan 2010–15
identifies weak human and institutional development as one
of the economy’s key binding constraints. Most public sector
departments are characterized by coordination failures, cor-
ruption, endemic malaise, and weak institutional linkages
among relevant stakeholders, including the Ministry of
Finance Planning and Economic Development, sector line
ministries, and the private sector. As a result, budgeting
processes are inefficient, and scarce investment resources are
not allocated rationally. Problems are particularly acute at
the local government level. The central government needs to
address these problems as it devolves planning, delivery, and
management of basic services to local governments (EPRC
2010). There is a need to increase human and institutional
capacity to meet the growing needs of local governments as
the scope of their functions broadens.
CONCLUSION
Uganda has experienced relatively strong growth and
poverty reduction since the 1990s. The growth period was
characterized by sustained macroeconomic stabilization,
adjustment, and structural reform efforts, which had a large
impact on most sectors. Efforts included price liberaliza-
tion, financial sector liberalization, public enterprises
reform, and reform of the civil service.
Growth has been accompanied by rising inequality and
very high unemployment levels, however, especially among
youth. Moreover, growth has been registered in sectors
(particularly services) whose contribution to employment
is limited. More needs to be done to translate impressive
performance at the macro level into improvements in
the welfare of the majority of Uganda’s people, especially
the rural poor.
Looking forward, Uganda needs to address several issues
in order to enjoy equitable growth. First, policy makers need
to increase the pace of transformation of the agricultural
sector and strengthen the sector’s weak link to industry,
where outputs from agriculture can be used as inputs. Given
that the sector employs 70 percent of the workforce, doing so
would spur equitable growth and reduce high unemployment.
CHAPTER 2: BUILDING ON GROWTH IN UGANDA 59
 Second, policy makers have to find new ways to mobilize policy makers also need to address other key challenges,
domestic revenues to finance the budget. To address the including weak human and institutional capability, a weak
growing trade deficit, Uganda should continue diversifying private sector, poor infrastructure, poor governance, and
its export base into higher-value-added products. Third, high population growth rates.

ANNEX 2.A

Table 2.A1 GDP in Uganda, by Economic Activity, 2001/02–2008/09

(percent, at constant 2002 prices)
Item 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10
GDP at market prices 8.6 6.6 6.8 6.3 10.8 8.4 9.0 7.2 5.8
Agriculture, forestry, and fishing 7.1 2.1 1.6 2.0 0.5 0.1 1.3 2.5 2.1
Cash crops 12.5 3.2 7.3 –5.5 –10.6 5.4 9.0 5.6 –2.9
Food crops 5.7 2.2 –1.5 –0.2 –0.1 –0.9 2.4 2.6 2.7
Livestock 4.0 3.5 4.7 3.0 1.6 3.0 3.0 3.0 3.0
Forestry 6.8 5.2 3.1 6.5 4.1 2.0 2.8 6.3 2.4
Fishing 13.8 –4.3 9.6 13.5 5.6 –3.0 –11.8 –0.7 2.6
Industry 7.4 9.5 8.0 11.6 14.7 9.6 9.1 5.8 8.9
Mining and quarrying 12.2 12.8 1.7 27.2 6.1 19.4 3.0 4.3 12.8
Manufacturing 6.7 4.4 6.3 9.5 7.3 5.6 7.6 10.0 5.9
Formal 7.7 4.6 8.3 11.8 7.8 4.9 9.2 12.0 6.1
Informal 4.5 4.0 1.7 3.6 6.0 7.7 3.3 4.4 5.5
Electricity supply –1.7 3.7 7.7 2.1 –6.5 –4.0 5.4 10.6 8.9
Water supply 3.0 3.9 4.2 3.9 2.4 3.5 3.8 5.7 2.7
Construction 10.1 14.6 10.0 14.9 23.2 13.2 10.8 3.7 10.9
Services 11.0 7.4 7.9 6.2 12.2 8.0 10.2 8.8 5.8
Wholesale and retail trade; repairs 7.4 5.1 6.3 7.2 12.3 10.4 14.7 9.7 –0.3
Hotels and restaurants 2.8 8.2 9.5 6.5 8.7 11.3 10.7 4.5 4.5
Transport and communications 17.8 14.9 15.8 9.8 17.1 17.7 21.3 14.3 15.1
Road, rail, and water transport 6.0 5.6 8.9 6.7 12.8 9.5 20.8 12.9 2.8
Air transport and support 0.5 5.8 13.8 19.4 6.9 13.8 17.8 –3.6 –1.2
Posts and telecommunication 76.5 40.4 28.6 11.8 26.2 29.1 22.6 19.8 30.3
Financial services 32.6 13.2 0.0 13.0 31.7 –11.9 17.1 25.4 21.1
Real estate activities 5.4 5.5 5.5 5.5 5.6 5.6 5.6 5.7 5.7
Other business services 11.6 7.6 7.0 9.2 12.5 8.0 10.8 12.4 10.4
Public administration and defense 20.4 3.6 7.7 –5.4 15.8 –6.3 12.1 5.5 3.9
Education 14.2 7.2 9.1 4.4 9.4 10.6 –6.5 4.3 –0.5
Health 18.0 13.7 0.9 5.6 12.9 2.7 –4.8 –3.2 11.0
Other personal services 8.5 8.5 16.1 15.0 14.1 13.4 12.8 12.3 11.8

Source: Bank of Uganda 2009.

Table 2.A2 Composition of Uganda’s Exports, 1997/98–2008/09

(percent)
Year Coffee exports Noncoffee exports Informal cross-border exports
1997/98 58.7 41.3 0.0
1998/99 55.9 44.1 0.0
1999/2000 42.0 56.6 1.3
2000/01 24.5 75.5 0.0
2001/02 18.1 81.9 0.0
2002/03 20.8 77.9 1.3
2003/04 17.0 78.4 4.6
2004/05 16.3 71.4 12.3
2005/06 16.6 67.3 16.0
2006/07 15.2 67.4 17.4
2007/08 13.4 45.4 41.2
2008/09 10.9 38.9 50.2

Source: Bank of Uganda.

60 CHAPTER 2: BUILDING ON GROWTH IN UGANDA

 Table 2.A3 Poverty Headcount, Poverty Gap, Poverty Severity, and Income Inequality Estimates, 1992/93–2005/06
Poverty headcount Poverty gap Severity of poverty Gini coefficient
Region 1992/93 2002/03 2005/06 2009/10 1992/93 2002/03 2005/06 2009/10 1992/93 2002/03 2005/06 2009/10 1992/93 2002/03 2005/06 2009/10
National 56.4 38.8 31.1 24.5 20.9 11.9 8.7 6.8 10.3 5.1 3.5 2.8 0.365 0.428 0.408 0.426
Rural 60.3 42.7 34.2 27.2 22.6 13.1 9.7 7.6 11.2 5.7 3.9 3.1 0.328 0.363 0.363 0.375
Urban 28.8 14.4 13.7 9.1 8.7 3.9 3.5 1.8 3.7 1.6 1.4 0.6 0.396 0.483 0.432 0.447
Central 45.6 22.3 16.4 10.7 15.3 5.5 3.6 2.4 7 1.9 1.3 0.8 0.395 0.46 0.417 0.451
East 58.8 46 35.9 24.3 22 14.1 9.1 5.8 10.9 6 3.4 2.1 0.327 0.365 0.354 0.319
North 73.5 63 60.7 46.2 30.3 23.4 20.7 15.5 15.8 11.5 9.2 7.3 0.345 0.35 0.331 0.367
West 52.7 32.9 20.5 21.8 18.7 8.5 5.1 5.4 9.0 3.3 1.8 2.0 0.319 0.359 0.342 0.375

Source: Ssewanyana and Okidi 2007.

Note: The poverty headcount is the percentage of people estimated to be living in households with real private consumption per adult equivalent below the poverty line for their region. The
poverty gap is the sum over all individuals of the shortfall of their real private consumption per adult equivalent and the poverty line divided by the poverty line. Severity of poverty is the
squared poverty gap (the sum over all individuals of the square of the shortfall of their real private consumption per adult equivalent and the poverty line divided by the poverty line).
61
 Table 2.A4 Targets for and Status of Progress toward Millennium Development Goals and Poverty Eradication
Action Plan (PEAP)
Target possible
with better
Target policies,
2007/08 2013/14 2015 possible at institutions,
PEAP PEAP MDG current and additional
MDG/target/indicator 1990 2005/06 2009/10 target target target trend? funding?
1. Eradicate extreme hunger and poverty:
Halve, between 1990 and 2015, the
proportion of people whose income is
less than $1 a day
1.1 Poverty headcount ratio (percent) 56 31.1 24.5 n.a. 28 28 Met Yes
1.2 Prevalence of child malnutrition
(percent of children under five) 23 23 n.a. n.a. n.a. 12 No Yes
2. Achieve universal primary education:
Ensure that, by 2015, children
everywhere, boys and girls alike, will be
able to complete a full course of
primary schooling
2.1 Net primary enrollment ratio
(percent of children 6-12)
Boys 58 84 82.4 90 100 100 Yes Yes
Girls 48 85 83.2 89 100 100 Yes Yes
2.2 Primary completion rate
(percent of boys and girls) n.a. 56 n.a. 69 n.a. 100 No Yes
3. Promote gender equality and empower
women: Eliminate gender disparity in
primary and secondary education,
preferably by 2005, and in all levels of
education no later than 2015
3.1 Ratio of girls to boys in primary
education (percent) 83 99 n.a. 100 100 100 Met Yes
4. Reduce child mortality: Reduce by
two-thirds, between 1990 and
2015, the under-five mortality rate
4.1. Under-five mortality rates
(per 1,000) 177 152 n.a. n.a. n.a. 53 No Uncertain
4.2 Infant mortality rate
(per 1,000 live births) 98 88 n.a. 68 n.a. 32 No Uncertain
4.3. Immunization against DPT
(percent of all children) 45 83 n.a. 90 n.a. n.a. n.a.
5. Improve maternal health: Reduce
by two-thirds, between 1990 and 2015,
the under-five mortality rate
5.1 Maternal mortality ratio (modeled
estimate, per 100,000 live births) n.a. 505 n.a. 354 n.a. 126 No Uncertain
5.2 Deliveries in health care centers
(percent of all deliveries) n.a. 24 n.a. 50 n.a. n.a. Met Yes
6. Combat HIV/AIDS, malaria and
other diseases: Have halted by 2015
and begun to reverse the spread of
HIV/AIDS
6.1 Prevalence of HIV, total
(percent of adult population) 20 6.2 n.a. 5 n.a. <20 Met Yes
7. Ensure environmental sustainability:
Integrate the principles of sustainable
development into country policies
and programs and reverse the
loss of environmental resources

(continued next page)

62 CHAPTER 2: BUILDING ON GROWTH IN UGANDA

 Table 2.A4 (continued)
Target possible
with better
Target policies,
2007/08 2013/14 2015 possible at institutions,
PEAP PEAP MDG current and additional
MDG/target/indicator 1990 2005/06 2009/10 target target target trend? funding?
7.1 Forest area (percent of total
land area) n.a. 24 n.a. 27 30 >24
7.2 Access to safe water
(percent of population)
Urban 45 86.8 92.3 100 n.a. 90 Yes Yes
Rural 45 63.6 69.5 90 n.a. 90 Yes Yes
7.3 Access to improved sanitation
(percent of population)
Urban n.a. 65 n.a. 100 n.a.
Rural n.a. 55 n.a. 80 n.a.
7.4 Tilled land (percent of land) n.a. 13 n.a. 17 25
8. Develop a global partnership for
development
8.1 Debt service (percent exports
of goods and services) 305 238 187 Yes Yes

Source: 2004 PEAP; Uganda Bureau of Statistics various years.

Note: n.a. is not applicable. PEAP is Poverty Eradication Action Plan.

NOTES
1. Some reversals in the decline occurred during this period:
the share of poor people rose from 34 percent in 1999/00 to
38 percent in 2002/03 (Appleton and Ssewanyana 2004).
2. The program targeted relaxation of marketing infra-
structure constraints, production and dissemination of
technology, removal of financial constraints, improvement
of land tenure and policy, formation of farmers organiza-
tions, improvement of human resources and information,
promotion of on-farm and off-farm storage, and efforts to
reduce environmental degradation and mitigate the effects
of HIV/AIDS.
3. Electricity coverage stood at 8 percent of the population
in 2009, far below coverage in other low-income countries,
which stands at 35 percent. At 190 meters per square kilo-
meter, the density of roads in Uganda is much higher than
in the average low-income country (where the average is
126 meters per square kilometer), but the roads are of poor
quality, with only 4 percent of the road network paved. Only
26 percent of the railway network in Uganda is functional.
Access to improved water stands at 51 percent of the
population, far below the 64 percent for the average of a
low-income country (Republic of Uganda 2010).
4. Taxing the informal sector could be implemented by
taxing small establishments based on characteristics such
as turnover, profits, and location of business.
BIBLIOGRAPHY
AfDB (African Development Bank). 2009. “Uganda Eco-
nomic Outlook 2009/10.” Africa Economic Outlook
2009/10. Tunis.
Appleton, S., and S. Ssewanyana. 2004. “Poverty Estimates
from the Uganda National Household Survey of
2002/03.” Economic Policy Research Centre, Kampala.
Bank of Uganda. 2009. Annual Report 2008/09. Kampala.
Bevan, D., C. Adam, J. Okidi, and F. Muhumuza. 2003. “Eco-
nomic Growth, Investment and Export Promotion.”
Paper prepared for Uganda’s Poverty Eradication Action
Plan Revision, 2002/03.
Collier, P., and R. Reinikka. 2001. “Reconstruction and Lib-
eralization: An Overview.” In Uganda’s Recovery, The
Role of Farms, Firms and Government, ed. R. Reinikka
and P. Collier, 15–48. Washington, DC: World Bank.
Dijkstra, A. G., and J. K. Van Donge. 2001. “What Does the
‘Show Case’ Show? Evidence of and Lessons from Adjust-
ment in Uganda.” World Development 29 (5): 841–63.
EPRC (Economic Policy Research Centre). 2010. Corrup-
tion Trends in Uganda: Using Data Tracking Mechanism.
Kampala.
Kappel, R., J. Lay, and S. Steiner. 2005. “Uganda: No More
Pro-poor Growth?” Development Policy Review 23 (1):
27–53.

CHAPTER 2: BUILDING ON GROWTH IN UGANDA 63

 MFPED (Ministry of Finance Planning and Economic
Development). 2009. “Budget Speech for Financial Year
2009/2010.” Delivered at the Fourth Session of the Eighth
Parliament of Uganda, June 11.
Okidi, J. A., S. Ssewanyana, L. Bategeka, and F. Muhumuza.
2007. “Uganda’s Experience with Operationalising Pro-
Poor Growth, 1992–2003.” In Delivering on the Promise of
Pro-Poor Growth: Insights and Lessons from Country
Experiences, ed. T. Besley and L. J. Cord, 169–98. Wash-
ington, DC: Palgrave Macmillan and World Bank.
Republic of Uganda. 2010. National Development Plan
2010/11–2014/15. Kampala.
Senoga, E., J. M. Matovu, and E. Twimukye. 2009. “Tax
Evasion and Widening the Tax Base in Uganda.” EPRC
Research Series 63, Economic Policy Research Centre,
Kampala.
Ssewanyana, S. 2009. “Growth, Inequality, Cash Transfers
and Poverty in Uganda.” Country Study 19, International
Policy Centre for Inclusive Growth, United Nations Devel-
opment Programme, Brasilia. http://www.ipc-undp.org/
pub/IPCCountryStudy19.pdf.
64 CHAPTER 2: BUILDING ON GROWTH IN UGANDA
———. 2010. “Combating Chronic Poverty in Uganda:
Towards a New Strategy.” EPRC Research Series 67,
Economic Policy Research Centre, Kampala.
Ssewanyana, S., and J. A. Okidi. 2007. “Poverty Estimates
from the Uganda National Household Survey of
2005/06.” EPRC Occasional Paper 34, Economic Policy
Research Centre, Kampala.
Transparency International. 2010. www.transparency.org
Uganda Bureau of Statistics. 2009. Statistical Abstract. Kampala.
———. 2010. Uganda National Household Survey 2009/10:
Socio-Economic Module, Abridged Report. Kampala.
UNDP. 2010. Uganda MDG Progress Report, Kampala.
World Bank. 2007. “Uganda, Moving beyond Recovery:
Investment and Behavior Change for Growth.” Country
Economic Memorandum Volume II: Overview, Report
39221-UG, Africa Region, Poverty Reduction and Eco-
nomic Management Unit, Washington, DC.
———. 2009. World Development Indicators. 2009. Wash-
ington, DC.
———. 2010. World Development Indicators. 2010. Wash-
ington, DC.
 CHAPTER 3
Rapid Growth and Economic
Transformation in Mozambique,
1993–2009
Antonio M. D. Nucifora and Luiz A. Pereira da Silva
ozambique’s economic growth after 1992 and the
M
poverty alleviation achieved since then constitute
an extremely successful development take-off.
Average real GDP growth rate soared from 0 percent
(1981–1992) to 8.1 percent (1993–2008), making
Mozambique the fastest-growing non-oil economy in Sub-
Saharan Africa over the period (AfDB 2009) (figure 3.1). As
a result, per capita GDP has doubled since 1992 (figure 3.2).
The poverty headcount dropped rapidly during the early
years of this process, falling from 69 percent in 1996 to
54 percent in 2002.
The political stability brought about by the end of the
armed conflict in 1992, a first wave of structural reforms
coinciding with responsible macroeconomic policies, and
the support of the international donor community have
been the bedrock of Mozambique’s transformation.
They have enabled the country to better face the develop-
ment challenges that lie ahead—in particular, how to tap
into the country’s resources (labor and land) and the entre-
preneurial drive of its people to promote strong, sustain-
able, and inclusive economic growth. Large mineral
resources and the recent growth in extraction activities is
allowing Mozambique to expand further its development
strategy. On the one hand this new revenue source can
allow the country to become less dependent on foreign aid
and define with more independence new elements of its
development agenda. On the other, it raises a number of
textbook challenges such as how to trigger externalities
(e.g., job creation) in other sectors and how to effectively
employ these revenues, while improving governance and
accountability.
This chapter provides a brief overview of Mozambique’s
profound economic transformation and highlights the
major drivers of change and the challenges ahead. It does
not discuss the macroeconomic policies and performance
that underpinned the economic transformation over the
past two decades (for a detailed discussion of macroeco-
nomic and stabilization issues in Mozambique, see Clement
and Peiris 2008).
GROWTH DRIVERS AND ECONOMIC
TRANSFORMATION
Sound macroeconomic environment allowed donors to
contribute substantial amounts of aid to Mozambique,
averaging about 14 percent of GDP a year since 1993 (figure
3.3). That aid inflows have remained fairly constant as a per-
centage of GDP over the period, during a period when the
economy grew at an average annual rate of more than 8 per-
cent, reflects the rapid growth in aid inflows in U.S. dollar
terms. These large inflows financed investments in educa-
tion and health, as reflected in the rapid improvements in
human development indicators (table 3.1).1 They also
financed substantial investments in rebuilding the country’s
roads, ports, and railways, which had been shattered by
16 years of war.
65
 Figure 3.1 Real GDP Growth in Mozambique and Sub-Saharan Africa, 1981–2005

20

15

10

5
Percent
0

–5

–10

–15

–20

05
97

01
81

91

99

03
83

85

87

89

93

95

20
19

20
19

19

19

20
19

19

19

19

19

19
Mozambique Sub-Saharan Africa (median)

Source: World Bank World Development Indicators various years.

Figure 3.2 Per Capita GDP in Mozambique and Sub-Saharan Africa, 1980–2006

400

300

200
In constant (2000) US$

100

–100

–200

–300
80

82

84

86

88

90

92

94

96

98

00

02

04

06
19

19

19

19

19

19

19

19

19

19

20

20

20

20

Mozambique Sub-Saharan Africa (median)

Distance from median

Source: World Bank World Development Indicators various years.

Good macroeconomic management also attracted sub-
stantial foreign direct investment (FDI). FDI inflows
increased from an average of 1.5 percent of GDP in
1993–98 to an average of 5.2 percent of GDP in 1999–2010
(figure 3.4). In 2009 and 2010 FDI reached an estimated
$900 million, about 9 percent of GDP. A large part of these
inflows has funded large investment projects in the mining
sector, underpinning recent export performance in
Mozambique.
From a supply-side perspective, capital accumulation,
higher quality–adjusted labor input, and positive aggregate
productivity performance were important determinants of

66 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009

Figure 3.3 Gross Annual Aid Inflows to Mozambique, 1998–2010

2,000 20

Millions of dollars
15

Percent of GDP
1,500

1,000 10

500 5

0 98 0

99

00

01

02

03

04

05

06

07

08

09

te 0
jec 01
19

19

20

20

20

20

20

20

20

20

20

20

d)
ro 2
(p
Gross aid inflows (percent of GDP), righthand side
Gross aid inflows (USD million), lefthand side

Source: Bank of Mozambique, IMF, and World Bank.

Table 3.1 Selected Social Indicators for Mozambique, 1990–2008

Latest single year
Indicator 1990–1996 2000–03 2006–08
Primary school enrollment (net %) 43.000 56.000 96.000
Primary school enrollment (gross %) 61.000 84.000 105.000
Ratio of girls to boys in primary
and secondary education (%) 72.000 75.000 85.000
Under-5 mortality rate
(per 1,000 live births) 212.000 178.000 138.000
Infant mortality rate
(per 1,000 live births) 145.000 122.000 93.000
Life expectancy at birth (years) 27.000 42.000 42.000
Physicians per 1,000 people 0.012 0.024 0.030
Inmunization, DPT (% of children
under 12 months) 60.000 72.000 72.000
Inmunization, measles (% of
children under 12 months) 58.000 77.000 77.000
Access to improved water
sources (% of population) 39.000 42.000 48.000
Access to sanitation facilities
(% of population) 22.000 27.000 31.000
Source: World Bank staff, based on data from Instituto National de Estatistica.
Note: Latest single year means the latest year for which data are available in a given period.

growth in Mozambique. Growth accounting exercises indi-
cate that physical investments partly associated with mega-
projects and significant improvements in education led to
growth dynamics heavily influenced by the accumulation of
human (quality-adjusted) and physical capital between
1993 and 2008. Figure 3.5 shows the average capital contri-
bution to growth strengthening from 1999 onward, a result
consistent with the timing of mega-projects. Aggregate pro-
ductivity growth (total factor productivity)—measured as
the residual of output growth after labor and capital contri-
butions are subtracted—was also important. Figure 3.6 sug-
gests roughly balanced growth of inputs and productivity.2
Productivity increases contributed about a third to overall
growth between 1993 and 2008. These results are broadly
consistent with those of Jones (2006) and Vitek (2010).3
Jones presents evidence suggesting a positive contribution
(crowding in) of public capital accumulation in the after-
math of the civil war, which opened the door for private
investment to play an increasingly important role in
Mozambique’s growth.

CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 67

 Figure 3.4 Foreign Direct Investment in Mozambique, 1993–2010

1,000 10

800 8

Millions of dollars

Percent of GDP
600 6

400 4

200 2

0 0
93
94
95
96
97
98
99
00
01
02
03
04
05
06
07
08

ro 2 9
te 0
0
jec 01
19
19
19
19
19
19
19
20
20
20
20
20
20
20
20
20
20

d)
FDI (Percent of GDP), right hand side

(p
FDI (millions of dollars), left hand side

Source: Bank of Mozambique.

Figure 3.5 Decomposition of Growth in Mozambique by Factors, 1993–2008

14
12
10
8
6
Percent

4
2
0
–2
–4
–6
93

94

95

96

97

98

99

00

01

02

03

04

05

06

07

08
19

19

19

19

19

19

19

20

20

20

20

20

20

20

20

20
Capital contribution Labor contribution (quality adjusted)
Total factor productivity GDP growth

Source: World Bank staff calculations based on World Development Indicators (World Bank various years).

Figure 3.6 Average Factor Contributions to Growth in A level change in aggregate investment from mega proj-
Mozambique, 1993–2008 ects triggered strong export and import activity from 1999
onward, with direct impact on aggregate demand. The
expenditure components of output (figure 3.7) indicate
that GDP and consumption have been trending up since
the end of the civil war but that growth after 1999 has been
32%
different. Strong investment performance starting in
45%
1999—which matured into strong export performance
from 2001 onward—allowed for continued increases in
23% output and consumption at the same time that imports
were growing. This pattern inaugurated a different growth
model for Mozambique, whereby strong exports, though
Capital Labor (quality-adjusted) offset by accompanying imports, reflected stronger private
Total factor productivity sector activity.
Source: World Bank staff calculations based on World Development Indica- More recently, as a result of sharp increases in the price of
tors (World Bank various years). aluminum from 2004 onward, the pace of profit repatriation

68 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009

Figure 3.7 Components of GDP Expenditure in Mozambique, 1993–2006

160,000

140,000

120,000

Millions in local currency

100,000

80,000

60,000

40,000

20,000

0
93

94

95

96

97

98

99

00

01

02

03

04

05

06
19

19

19

19

19

19

19

20

20

20

20

20

20

20
Consumption Investment Exports Imports GDP

Source: World Bank 2010a.

Figure 3.8 Gap between GDP and GNP in Mozambique, 1994–2007

14 180,000

160,000
12
140,000

Millions in local currency

10
120,000
Growth rate (%)

8 100,000

6 80,000

60,000
4
40,000
2
20,000

0 0
00

01
99

06
03

05
02

04

08
07
98
93

94

97
96
95

20

20
19

20
20

20
20

20

20
20
19
19

19

19
19
19

GDP growth rate (left hand side) GNP growth rate (left hand side)
GDP level (right hand side) GNP level (right hand side)

Source: World Bank 2010a.

associated with mega-projects has widened the gap between abundant factors of production, particularly labor, land, and
GNP and GDP (figure 3.8). This stylized fact is a reminder entrepreneurial drive.
of a key challenge for the country: the need to compliment Although the private sector has spearheaded large mega-
the current export drive based on mega–projects with a projects in the mining sector, its role in economic growth
diversification strategy that activates a broader set of growth over the past decade should not be overemphasized.
drivers and taps more effectively into Mozambique’s more Sonne-Schmidt, Arndt, and Magaua (2008) developed a

CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 69

 methodology to evaluate the direct contribution of three
key operational mega-projects (Mozal, Sasol, and Moma)
to growth at factor cost. They conclude that the direct
effects of these mega-projects were an increase in the
average annual growth rate between 1996 and 2006 of
about 1 percentage point (average annual growth rate was
about 8 percent). Moreover, as foreign-owned, capital-
intensive, export-oriented companies benefiting from
fiscal exemptions, mega-projects made only a small contribu-
tion to job creation, tax revenue, use of domestic intermedi-
ate inputs, and profit reinvestment in Mozambique.
Benito-Spinetto and Moll (2005) present similar results.
Using a computable general equilibrium (CGE) model to
replicate the performance of the Mozambican economy,
they find that about half of the 8 percent average GDP
growth during 1994–2004 was driven by catch-up growth in
the agricultural sector (as people returned to the fields fol-
lowing the end of the war), a quarter was driven by the
growth in aid inflows (and the resulting investments in
social and physical infrastructure sectors), and a quarter
was equally divided between mega-projects and growth in
other private sector activities. Clearly, the initial bet on
mega-projects, despite the usual uncertainty in quantifying
their precise economic effects, rested more on their capacity
to generate immediate revenue rather than truly create
downstream externalities (especially jobs and demand for
industrial and services inputs).
Figure 3.9 Sectoral Contribution to Growth in Mozambique, 1994–2006
20
15
10
Percent
–5
94
95
96
97
19
19
19
19
Agriculture
Construction
Source: World Bank various years.
70 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009
Indeed, the positive total factor productivity reflects
the reshuffling of labor resources and a rapid transforma-
tion in the output structure of the economy. In a postwar
context—where reconstruction efforts, a first wave of
structural reforms, and a stabilized macroeconomic envi-
ronment spurred private sector activity—the Mozambican
economy posted double-digit growth rates in mining,
manufacturing, construction, electricity, gas, and water.
Sectoral output shares experienced substantial changes
between 1993 and 2006, especially in agriculture, which
nearly halved its output participation, despite the positive
contributions of agriculture to overall economic growth
(figures 3.9 and 3.10). Sectoral output reflects a deep
transformation to more productive sectors that generated
positive composition effects in aggregate productivity.4
Nevertheless, although economic growth has been strong
and has underpinned the process of economic transfor-
mation, the structure of the economy still remains nar-
rowly based on subsistence agriculture and a few isolated,
albeit large, mega-projects. Agriculture, which employed
about 78 percent of the economically active population,
accounted for 25 percent of GDP in 2009. It was followed
by manufacturing (12 percent, two-thirds of which was
accounted for by one large aluminum smelter); trade and
retail services (11 percent); transport and communica-
tions (10 percent); financial services (7 percent); and
extractive industries (1 percent). Going forward, a key
98
99
00
01
02
03
04
05
06
19
19
20
20
20
20
20
20
20
Mining Manufacturing Electricity, gas, and water
Services GDP growth
Figure 3.10 Changes in Sectoral Output Share in Mozambique, 1993–2008

45 45
40 40

Real growth rate (percent)

Sectoral share (percent)
35 35
30 30
25 25
20 20
15 15
10 10
5 5
0 0
re

w and

g
rin

in
io
tu

in
er
ct
tu
ul

M
as
ru

at
ac
ric

,g
st
uf
Ag

on

ity
an

ic
C
M

tr
ec
El
1993 Share 2008 Share Average annual real growth (93–08)

Source: World Bank various years.

challenge is not only to support the continuation of struction, and transportation sectors between 2003 and
intersectoral (structural) transformations but to ensure 2006 (table 3.3).6 The dominance of large companies in for-
within-sector productivity gains by addressing impedi- mal employment generation reflects the lack of growth in
ments to microeconomic efficiency (World Bank 2008a). small and medium enterprises. The “entry margin” (cre-
Economic transformation in Mozambique entailed a ation of formal jobs through new firms) also appears con-
profound rebalancing of sectoral roles, with the (pri- siderably more stifled for micro and small firms than for
vate) service sector absorbing labor at a particularly medium and large firms, perhaps a reflection of the fact that
strong pace. Agriculture displays the lowest level of labor entry costs and overall business environment constraints
productivity of all sectors (figure 3.11). Between 1996 impinge more severely on smaller producers.
and 2003 its output contribution to GDP fell by 6 per- The concentration of the economy in a few large compa-
centage points (to 27 percent), and its share of employ- nies is also reflected in export patterns. The export basket
ment in the total labor force fell by 8 percentage points remains extremely limited, with only 15 products register-
(to 82 percent) (figure 3.12). Although output shares ing exports of more than $1 million a year. More than half
were gained primarily by the industry sector, employ- of Mozambique’s exports remain concentrated in the Mozal
ment shares were gained by the private service sector. aluminum smeltering enclave (figure 3.13), virtually all of
While the increase in the employment share of the pri- whose production goes to the European Union. Electricity
vate service sector has benefited the poor (World Bank and gas are also important export products (bought largely
2008b),5 the challenge to generate an even greater number by South Africa).
of jobs at higher levels of the value chain remains. The Overall, Mozambique’s exports increased significantly
overwhelming concentration of employment (more than (in real terms), with average annual growth rates of
78 percent) remains in agriculture. In urban areas, retail 14.9 percent over 1995–99 and 19.8 percent over 2000–08.
trade and vehicle repair employ a quarter of all workers; Most of this growth, however, was driven by exports from
manufacturing industries employ 7.5 percent (table 3.2). the mega-projects; although excluding mega-projects,
Formal employment generation has been very limited in exports grew at an average rate of 10 percent a year during
recent years and is confined to a few large companies. Estab- 2000–08, which represents a reasonable performance.
lishment-level data from a sample of 600 firms suggest that Progress in penetrating Southern Africa Development
large firms (more than 100 employees) and some medium- Community (SADC) countries continues; Mozambique
large firms (about 100 employees) were responsible for the should position itself to exploit non–SADC markets as
majority of jobs created in the manufacturing, services, con- well. Mozambique exploits only 2 percent of the potential

CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 71

 Figure 3.11 Labor Productivity in Mozambique by Sector (Relative to Agriculture), 1996 and 2003

30

Average labor productivity normalized

25

20

to agriculture
15

10

0
Industry Services (private) Services (public)

1996 2003

Source: World Bank 2008b.

Note: Figures are normalized to agriculture.

Figure 3.12 Changes in Employment versus Output Shares in Mozambique, by Sector, 1996–2003

100

80
Change in output share between

60
1996 and 2003 (percent)

40

20

–20

–40

–60
–50 0 50 100 150 200
Change in employment share between 1996 and 2003 (percent)

Industry Services (public) Services (private) Agriculture

Source: World Bank 2008b.

Note: Services (private) include trade, and transport. Services (public) include health, education, and public administration. The larger the bubble, the
higher the labor productivity.

bilateral flows for all the products it exports (Gillson A STRATEGY FOR BROADLY SHARED
2008).7 Although this is comparable to other countries in ECONOMIC GROWTH
the region, it is far lower than the levels achieved by more
Rapid economic growth was accompanied by significant
advanced and better integrated economies (such as South
strides in reducing poverty during the stabilization phase,
Africa) (figure 3.14).8
after the peace accords and up to 2003.9 Household survey

72 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009

 Table 3.2 Distribution of Employment in Mozambique, by Sector, Gender,
and Location, 2004/05
(percent)
Sector Total Men Women Urban Rural
Agriculture and fishing 78.5 68.0 87.3 40.1 92.7
Extractive industries 0.3 0.6 0.0 0.4 0.2
Manufacturing industries 3.1 5.4 1.2 7.5 1.5
Electricity, water, and construction 1.2 2.6 0.1 3.5 0.4
Retail trade and vehicle repair 9.2 11.7 7.0 25.9 3.0
Transport and communication 0.8 1.6 0.1 2.6 0.1
Financial services and rental 2.9 3.9 2.0 9.5 0.5
Administration 1.7 2.9 0.7 5.3 0.4
Education 1.6 2.3 1.0 3.5 0.9
Health and social work 0.7 0.8 0.5 1.6 0.3
Source: INE 2006.

Table 3.3 Number of Full-Time Employees in Sample of 600 Firms in Mozambique, by Firm Size, 2003 and 2006
Growth rate in
Change in number of employees
number of
between 2003 and 2006
employees between
2003 and 2006 Number of Number of In firms created In firms created
Firm size (percent) employees in 2003 employees in 2006 before 2003 after 2003
Micro 63 408 666 190 68
Small 35 2,064 2,803 493 246
Medium 36 4,630 6,323 992 701
Large 60 2,922 4,685 772 991
Total 44 10,024 14,477 2,447 2,006
Source: World Bank 2010b.
Note: Full-time employment in 2003 was used to determine firm size. For firms created after 2003, employment in 2006 was used to determine
firm size.

Figure 3.13 Mozambique’s Main Exports, 2008

1,600
1452
1,400

1,200
Millions of dollars

1,000

800

600

400
226 193
200 152
71 62 58 37 33 31 26 5 5 3 2
0
um

To y

as

ts

ts

s
um

e
oi
ed

na
ga

Te
t

aiz
rim

to
cc

uc

nu
ici

G
Su

ut
in

ni

na
se
ot
ba

M
tr

od
sh
um

ta
w

on
Ba
C
ec

e
he
pr

Ti
m
en

oc
Al

El

as

sa
d
oz

C
oo

Se
Fr

CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 73

 Figure 3.14 Mozambique’s Export Market Penetration, by Country, 2006

90
80.5
80
70
60

Percent
50
40
30
20.4
20
9.3
10
0.4 0.5 0.6 0.7 1.7 2.2
0
sia

ia

na

on

ca
w
di
tio

lan
ys

ri
hi

ala
ne

ni
In
ala

Af
ra

C
er

U
do

M
de

itz
M

h
an
In

ut
Fe

Sw

pe

So
ian

ro
Eu
ss
Ru

Source: UN COMTRADE (various years); Gillson 2008.

data indicate that the national poverty headcount fell
rapidly, from 69 percent in 1996 to 54 percent in 2003
(MPD 2004). Reduction in rural poverty was even more
pronounced, declining from 71 percent to 55 percent during
the same period.
The pace of poverty reduction appears to be slowing,
however, now that the post-war reconstruction effort and
the agriculture catch-up has been exhausted, and the exist-
ing pattern of economic growth is not generating a suffi-
cient number of jobs. The preliminary results of the 2009
nationally representative household budget survey are
mixed. Although there was significant progress on many
nonmonetary poverty indicators (such as access to educa-
tion and health services, increase in asset ownership by
households, and improvements in housing quality), the
poverty headcount suggests that poverty may have stag-
nated at about 55 percent during 2003 and 2009 (MPD
2010). Urban poverty continued to decline, although at a
much slower rate, reaching 50 percent, but rural poverty
increased to 57 percent.10 Although the food and fuel crisis
played a role in this outcome, the stagnation of poverty
reduction while the economy continued to grow at high
rates suggests that growth has become less inclusive in the
past few years.
These results are important because they seem to cor-
roborate what many academics in Mozambique (see, among
others, Castel-Branco 2002; Castel-Branco and Ossemane
2009) have been arguing for a number of years: that even
though GDP growth rates have been very high in Mozam-
bique, the current development strategy is lagging behind
expected promises. The model—which focuses on main-
taining macroeconomic stability, making public invest-
ments in infrastructure and increasing access to public
services (education, health, water, and electricity), and
attracting capital-intensive mega-projects (which are capital
intensive)—is still not bringing the increase in jobs and pro-
ductivity required to set the economy on a more diversified
growth pattern and eventually to reduce poverty by more
than under the current strategy. In addition, lack of progress
on the good governance agenda is putting strain on
Mozambique’s relationship with some of its development
partners, and raising further questions about the sustain-
ability of Mozambique’s development strategy (box 3.1).11
A more inclusive, labor-intensive economic growth strat-
egy is needed to reduce poverty and improve living stan-
dards (World Bank 2010a). Diversifying the economy into
labor-intensive sectors (including agriculture, agroprocess-
ing, manufacturing, and tourism) and increasing the com-
petitiveness of domestic production to replace imports and
diversify exports requires a new strategy focused on elimi-
nating barriers to private investment (both domestic and
foreign). Such a strategy would attempt to reduce excessive
regulation; simplify the trade and tax regime; reduce the
costs of hiring and firing workers; improve labor force skills;
eliminate the rigidity in the land tenure system (freeing up
access to land use rights and allowing their tradability);
strengthen transport logistics and facilitate better services/
articulation for exports; address the lack of standards; and
maintain sound fiscal, monetary, and exchange rate policies.
A number of constraints have to be addressed through a

74 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009

 Box 3.1 Signs of Political Maturity: Is Mozambique’s “Aid-Supported Growth Engine” Beginning to Shift?
Aid flows needed to finance public investments and
expand service delivery are becoming increasingly
uncertain in Mozambique but also for other Low Income
Countries. This is due, in part, from the budgetary con-
sequences of the 2008 global financial crisis in ODA allo-
cation in many OECD countries.
OECD donors have been considering Mozambique
an emblematic success story since the 1990s, especially
since the end of the Civil War (1992). Official develop-
ment assistance flows were significant and relatively
stable both for projects and for direct budget support,
given the country’s good performance and “donor-
agencies” high ratings. Throughout the early 2000s
changes occurred in Mozambican politics: from a rela-
tively strong showing in the first, post–Civil War elec-
tions in 1994, RENAMO’s electoral results declined
subsequently and posed a lesser threat to FRELIMO-
dominated politics. Changes also affected the FRELIMO
leadership, rotating toward a more business-oriented
framework. With more investment opportunities
brought by macroeconomic stabilization, positive
donor sentiment, and increased aid inflows, among
other things, a shift toward a market-oriented frame-
work took place, accentuated by the discovery in the
mid-2000s of significant mineral resources that trig-
gered private sector interest and investment. In recent
years, the discovery of such sizable reserves of miner-
als began to engender business interest in the country
by global players, triggering expectations of large,
mix of simultaneous policy reforms, investments, and
strengthening of institutions. They can be categorized as
business environment, factor markets, trade logistics, and
export-supporting institutions.
Aggressive reforms are required to improve the business
environment and entice potential investors. First, the regula-
tory environment is very unfriendly to small and medium
enterprises, and licensing, inspections, and red tape are a
heavy burden. This is arguably the most important priority
needing attention to unleash private sector potential, starting
with domestic investors. Second, the Mozambican tax and
tariff system is characterized by many different exemptions
whose cumulative effects are unclear. Additionaly, small
entrepreneurs normally do not benefit from current exemp-
tions, because they cannot afford the fixed cost involved in
CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009
private sector-led investments in these areas but
also in the accompanying infrastructure for export
corridors.
Inevitably these prospects brought more attention
to issues of economic governance such as the relation-
ship between an emerging class of businesses and
entrepreneurs with the dominant political elite. Fol-
lowing several years of unhappiness about mixed
progress in the good governance agenda and fears that
increasing growth of non-aid revenue would allow
Mozambique to become less aid-dependent. Some
OECD donors in late 2009 threatened to “cut budget
support” if very specific actions on political reforms
were not urgently undertaken. While the 2009–10
standoff has since been largely resolved, some of the
reasons and the political dynamics that led to the
impasse appear to signal a structural change in
Mozambique’s relationship with the budget support
donors. Many donors, however, consider such interfer-
ence in the country's political affairs as unwarranted
and refrain from entering into this type of conditional-
ity. In fact, the government has also started to consider
different possible scenarios to reduce its political vul-
nerability to aid-dependence, ranging from the possibil-
ity of revising some contracts with some mega-projects
that benefit from generous tax incentives, or turning to
some of BRIC countries for support, in addition to the
need to expand the tax collection base and diversify
revenues, among other strategies.
applying for them. The standard statutory tax regime
(excluding incentives) continues to place a substantial bur-
den on investors, notably small domestic entrepreneurs. The
combination of a statutory tax regime with relatively high
tax rates with a generous system of tax incentives influences
investment decisions and the allocation of resources in a
way that is not immediately visible but distortionary in the
long run. One widely recognized distortion is that the past
tax regime encouraged industries that used more capital
rather than labor. Mozambique’s fiscal climate could be
significantly improved by simultaneously reducing fiscal
incentives and the number and level of tax rates without
compromising government revenues. Third, despite
progress in recent years, tax administration remains
another impediment to Mozambique’s competitiveness
75
 and productivity. A large number of the firms interviewed
in the recent Investment Climate Assessment (World Bank
2010b) reported that tax administration is a major or severe
obstacle to doing business. Reducing the number of taxes on
the books and simplifying compliance with tax legislation
would reduce transaction costs of businesses.
Factor markets—land, labor, and capital—need to allocate
resources to their most productive uses. The 1997 Land Law
is generally considered to be exemplary in Africa. However,
10 years after its approval, its implementation is weak, and the
reality on the ground is one of extensive extralegal land mar-
kets, multiple claims to the same pieces of land, and a lack of
guidance on how to protect or compensate customary and
good-faith occupation rights. This situation has led to exten-
sive land speculation and corruption, potentially leading to
land conflicts and landlessness in the future. In the short run,
the land situation acts as a deterrent to investors, particularly
in tourism and agribusiness. Access to land for productive
purposes needs to be improved. The new labor law approved
in 2007 has substantially improved labor regulations for new
investments. However, the labor market could and should be
allowed to work even better to attract more domestic and for-
eign investment and sustain high growth rates.
In addition, despite significant progress, education levels
in Mozambique are still very low and constitute a structural
constraint for job creation in the formal sector. Quality of
education is key for inclusive growth. Professional services
(such as engineering, auditing, legal, and medical services)
are expensive and in short supply in Mozambique; at the
same time, many highly skilled Mozambicans are emigrat-
ing because they can earn more money abroad. This
paradox of not having enough supply of skills (making pro-
fessional services very expensive) while at the same time not
having enough demand for them (leading to emigration)
needs to be analyzed carefully to come up with an appropri-
ate package of policy reforms. Inappropriate standards often
stifle demand for services in areas such as engineering and
accounting. Professional associations can play a key role in
creating, with government support, a framework for regula-
tion and appropriate standards, and also monitor quality
and skills development. Immigration rules for qualified spe-
cialists who can provide managerial and technical expertise
to Mozambican enterprises are still very restrictive, particu-
larly for companies providing professional services. In the
short run, a more liberal policy of granting work visas to
skilled foreigners would help to develop the market for pro-
fessional services in Mozambique and would enable
Mozambican companies to compete better with foreign
companies providing such services.
76 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009
Legislation for the financial sector has been evolving
continuously, and all banks were privatized in the 1990s.
However, capital markets remain very shallow. The banking
system is now sound and highly profitable. Exceptionally
high ratios of capital adequacy (17 percent on Basle I) and
liquidity (55.4 percent) of commercial banks raise a legiti-
mate question as to whether there is sufficient financial
intermediation for a fast-growing economy. Only a quarter
of all districts have any kind of banking service, housing
finance faces big constraints, and trade finance is still
underdeveloped.
Trade logistics remain a serious constraint. Mozam-
bique has the potential to be a major outlet for southern
Africa based on its strategic location and congestion in
Durban port, provided that it becomes more logistics
friendly. Maputo port is the closest route for the largest
mining and manufacturing region of South Africa, while
Beira and Nacala are the closest ports for the other neigh-
boring countries. Yet corridor traffic potential remains
mainly untapped. Traffic going to and from Gauteng and
Mpumalanga provinces equals at least 700 million tons. If
Maputo could capture 1 percent of this traffic, its total
throughput would be doubled and port revenues would be
increased by several tens of millions of dollars. Despite a
rather successful concession process, transit traffic at
Maputo port remains relatively low, and the port operates
at less than 30 percent of its capacity. Beira port has also
operated at less than 40 percent capacity for the past five
years. Mozambique’s main deficiencies for trade mostly
derive from logistics problems and low transport reliability.
Low ratings in the logistic performance index (LPI) (2.29)
and in the timeliness index (2.83) put Mozambican ports at
a disadvantage compared with the port of Durban (with an
LPI of 3.53 and a timeliness index of 3.78). The World
Bank’s 2008 Doing Business report highlights a web of pro-
cedures taking 27 days for exports and 38 days for imports.
Some large South African shippers are still reluctant to
shift their transport routes from Durban to Maputo because
they still perceive the business climate there as unpre-
dictable, including a higher incidence of bribe payments for
the port of Maputo than for Durban (World Bank 2010a).
Also, the poor integration of trucking services with the rest
of the subregion explains why South African shippers are
still reluctant to use Mozambican ports. High transport
unpredictability is closely linked to low traffic volumes. As
a result, shipping lines do not call directly in Maputo and
cargo has to be transshipped in Durban or Mombasa.
Physical investments and reforms of customs procedures
should be carried out simultaneously to reach a critical mass
in port transport volume. At that point, international ship-
ping lines will make direct calls, which will greatly improve
predictability. Constructing the “one-stop border post” in
Ressano Garcia for trade with South Africa is of highest pri-
ority and should be quickly emulated in other large border
posts. At the same time, all the customs clearance proce-
dures including at the dry ports have to be revised. Long
truck immobilization creates unnecessary delays and costs
for shippers and transporters. Capital dredging of the Beira
port is also very important, as is the reconstruction of the
road between Nampula City and the border of Malawi.
Strengthening specialized export-supporting institu-
tions that can assist small- and medium-size exporters is
important for product and market development. Needed
are a national system for standards and quality controls; an
ICT-based proactive trade information system; product ID
cards, and CITEs (centers supporting knowledge, innova-
tion, and technology transfer and services otherwise not
available). In addition, more “articulators” are needed to
connect Mozambican businesses and products with
regional and global markets. For a country where produc-
tion is highly fragmented, such articulation initiatives are
critical to secure scale benefits and knowledge transfer.
There are already a number of ongoing articulation efforts
in Mozambique through nongovernmental organizations,
producer associations, private sector firms, and government
agencies. But more needs to be done to get the benefits of
economies of scale.
CONCLUDING REMARKS
The government of Mozambique has consistently pursued
sound fiscal, monetary, and external policies, demonstrating
a clear understanding of the benefits of a solid underlying
macroeconomic framework for fostering further private
sector-led growth. Mozambique’s stable macroeconomic
environment has allowed the economy to undergo signifi-
cant transformation over the past two decades, in part sup-
ported by a stable source of ODA funding. Sectoral output
composition has witnessed significant changes, with the
participation of agriculture in output diminishing as more
productive sectors grow. Although employment is still over-
whelmingly concentrated in agriculture, job growth in the
private sector (mostly informal) has led to increased partici-
pation of private services in output. Nevertheless, Mozam-
bique’s current growth model is not creating enough jobs or
widening the productive base in order to create an economy
in which wealth is broadly shared. However, the strategy
implemented since the discovery and exploitation of large
CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009
mineral resources has produced positive results. The country
has achieved high growth rates, is getting significant
amounts of export revenue and has begun to reduce its aid-
dependency, especially vis-a-vis traditional Western donors.
However, the trickle-down effect of such a bet has not (yet)
produced the expected impact on local businesses and labor
markets. Moreover, uneven income distribution in urban
areas and rising food prices during the last global commod-
ity price cycle triggered protests.
Shifting growth dynamics toward a more inclusive
pattern based on economic diversification and increased
participation of small and medium enterprises will require
policy action. Businesses’ interest in expanding and initiat-
ing operations in Mozambique is palpable. This ongoing
diversification of productive activities is bound to spur
transformations in the economic landscape. Economic
diversification and increasing engagement of small and
medium enterprises leveraged by sustained efforts to partic-
ipate in world markets represent a strategic orientation in
the Mozambican growth model that is required to acceler-
ate growth and reduce poverty. Harnessing such transfor-
mations for economic development will require strategic
policy steps. By promoting competition and trade integra-
tion measures, the government should induce further
private sector demand for Mozambique’s factors of produc-
tion, diversifying the economy and achieving shared and
sustained economic growth.
A more balanced growth path, that would rely on local
business initiatives, is fundamentally about creating the
right incentives and necessary underlying conditions to
foster entrepreneurial drive. These conditions can be
achieved through policies to improve the business environ-
ment, the functioning of factor markets and land alloca-
tion, and the trade infrastructure, aiming to strengthen or
create the institutions necessary for more absorption of
labor, in particular helping small and medium enterprises
to succeed. Such a shift requires political commitment and
attention to new institutional bottlenecks and implementa-
tion capacity.
The bet that Mozambique has taken is showing some
positive results. It was perhaps, given its chronic aid-
dependency during most of its post-independence history,
the only possible way to depart from it. This bet is not
exempt from risks, and there are several indications that the
political leadership is fully aware of many of them. Losing
too soon significant portions of ODA would certainly be
harmful to both macroeconomic stability and the quality of
policies, although it should be acknowledged that the
country has fully owned and implemented the standard
77
 agenda aimed at stability. The benefits from the export
windfall could be captured by vested interests and not re-
invested into the rest of the economy. There could be a sub-
stitution of aid-dependency by an excessive reliance on the
business interests of large private groups.
Nevertheless, it is overall quite a success story, despite
these risks and perhaps because of these risks. To the extent
that the country understands these challenges and works to
address them, especially in the efficient allocation of wind-
falls to development purposes, chances are that Mozambique
could be a bright spot in terms of a successful strategy that
overcame a civil war, instabilities of all sorts, lack of initial
resources, lack of human capital, poverty and excessive
dependency on external assistance.
NOTES
1. In primary education (grades 1–7), the number of
children more than tripled, from about 1.3 million in
1992 to 4.2 million in 2008. Net enrollment rates in pri-
mary education doubled, from 45 percent in 1998 to
more than 96 percent in 2008, with rates for girls rising
from 40 percent to 93 percent. The number of primary
school teachers increased from 30,000 in 1992 to 73,900
in 2008. The gross enrollment rate in lower secondary
education (ES1) increased from 4.8 percent in 1998 to 28.0
percent in 2008. The gross enrollment rate for upper sec-
ondary school (ES2) increased from 1.3 percent in 2008 to
8.0 percent in 2008. Under-five mortality rates decreased
from 212 per 1,000 live births in 1996 to 178 in 2003 and
138 in 2008. Infant mortality decreased from 145 per 1,000
live births in 1996 to 122 in 2003 and 93 in 2008. Maternal
mortality fell from an estimated 1,000 per 100,000 live
births in the early 1990s to 408 in 2003 and 340 in 2007.
The share of the population with access to an improved
water source increased from 39 percent in 1995 to 48 per-
cent in 2008. Although the rate of HIV prevalence
remains very high, the capacity of the health system was
expanded to start providing free antiretroviral treatment
for HIV infection.
2. The methodology applied to obtain the decomposition
was based on a Cobb-Douglas production function with
capital and quality-adjusted labor as inputs. Shares of the
inputs were considered fixed at 0.4 and 0.6. Data from the
Instituto National de Estatistica (INE 1998, 2004) were used
to adjust for labor quality.
3. Jones (2006) provides a thorough discussion of growth
accounting in Mozambique between 1980 and 2004. He
considers alternative and more flexible functional forms
for the production function; discusses perils of the growth
accounting exercises with specific Mozambican data con-
straints in mind; and reviews results from other authors,
78 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009
such as Sulename (2001); Benito-Spinetto and Moll
(2005); IMF (2005); Tahari, Akitoby, and Aka (2004); and
Ndulu and O’Connell (2003). Vitek (2010) provides a
growth accounting exercise in Mozambique between 1990
and 2008.
4. The shrinking of the labor force in agriculture in favor
of other more productive sectors is a general pattern in the
development process of many other advancing economies
and should not be viewed as a negative phenomenon.
5. Job creation in the nonagricultural informal sector
between 1997 and 2003 was also an important determinant
of poverty reduction (World Bank 2008b). It provided eco-
nomic opportunities to the poor and led to economywide
productivity gains.
6. The enterprise survey of 600 firms was carried out in
2008 by the World Bank for the Investment Climate
Assessment (World Bank 2010b). Each firm in the sample
was visited to collect employment information on 2003
and 2006.
7. The index essentially compares each product Mozam-
bique exports with all importing countries of the product
from all world sources. By tracing all bilateral flows for
each exported good, Gillson (2008) proposes a measure that
captures the extent to which markets have been penetrated
by Mozambican products. For instance, if country A exports
good X only to country B, but good X is also imported by
countries C and D only from country E, then country A is
exploiting one of three (or one-third of) bilateral flows for
good X.
8. Market penetration measures disaggregated by largest
export partners indicate that Mozambique is already
exploiting 80 percent of its potential bilateral flows to
South Africa.
9. This section draws heavily on the recent Country Eco-
nomic Memorandum on “Reshaping Growth and Creating
Jobs through Trade and Regional Integration” (World Bank
2010a).
10. These results should be treated with caution, however.
Substantial underreporting of food consumption has been
identified in the 2008/09 household survey data, which
could affect both the poverty levels and their trend over
time. The data problem also affects the 1996/97 and 2002/03
household surveys, albeit to a different extent. Additional
analytical work is required to test the robustness of the
poverty analysis in Mozambique by correcting for the data
problems.
11. The considerable amounts of aid and the presence of a
great number of aid agencies led to the establishment of
donor coordination mechanisms that were considered best
international practices until recently. A number of donor-
government working groups were established during a
“golden era” for coordination that brought benefits to both
donors and the government—in terms of transaction
costs—and culminated with the adoption of the Paris and
Accra Declarations. In parallel, a distinction began
between donors providing direct budget support (such as
a regrouping into a group of nineteen OECD countries or
the G-19) and those involved in more traditional project
activities. While important donors (such as the United
States) concentrated efforts in specific sectors (like health
and clean water), the G-19 elaborated a memorandum of
understanding to define quasi-contractual rules for dis-
bursing budget support, considered as a sign of a more
mature relationship with recipient governments. In fact,
while Mozambique’s success led the country to benefit from
larger shares of aid under direct budget support (by the
G-19), this theoretically more “advanced” form of aid, ade-
quate for countries with a solid track record, also makes the
country more dependent on the political and budgetary
cycles of OECD donors (see box 3.1).
REFERENCES
AfDB (African Development Bank). 2009. African Develop-
ment Indicators 2009. Tunis.
Benito-Spinetto, M., and P. Moll. 2005. “Macroeconomic
Developments, Economic Growth and Consequences for
Poverty.” Background Paper for the 2005 Country Eco-
nomic Memorandum. World Bank and International
Monetary Fund, Washington, DC.
Castel-Branco, Carlos Nuno. 2002. “Mega projectos e
estratégia de desenvolvimento.” http://www.iese.ac.mz/
lib/cncb/Mega_projectos_Moz_texto.pdf.
Castel-Branco, Carlos Nuno, and Rogéio Ossemane. 2009.
“Crises cíclicas e dasfios de transformação do padrão de
crescimento económico em Moçambique.” In Desafios
para Moçambique 2010, ed. L. Brito, C. Castel-Branco,
S. Chichava, and A. Francisco, pp. 141–82. Maputo: Insti-
tuto de Estudos Sociais e Económicos (IESE).
Clement, Jean A. P., and Shanaka J. Peiris, eds. 2008. Post-
Stabilization Economics in Sub-Saharan Africa. Washing-
ton, DC: International Monetary Fund.
Gillson, Ian. 2008. “Mozambique’s Export Market Penetration.”
World Bank, Poverty Reduction and Economic Manage-
ment (PREM), Africa Region, Washington, DC.
IMF (International Monetary Fund). 2005. Republic of
Mozambique: Selected Issues and Statistical Appendix.
IMF Country Report 05/311, Washington, DC.
INE (Instituto National de Estatistica). 1998. Relatório final
do inquérito sobre orçamento familiar (IOF), 1996/7.
Maputo.
CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009
———. 2004. Relatório final do inquérito aos agregados famil-
iares sobre orçamento familiar (IAF), 2002/3. Maputo.
———. 2006. National Labor Force Survey 2004/5. Maputo.
———. 2010. Relatório final do inquérito sobre orçamento
familiar (IOF), 2008/9. Maputo.
Jones, Sam. 2006. Growth Accounting for Mozambique,
1980–2004. Discussion Paper 22E, Ministry of Planning
and Development, National Directorate of Studies and
Policy Analysis, Maputo.
MPD (Ministry of Planning and Development), Govern-
ment of Mozambique. 2004. Second National Poverty
Assessment. Maputo.
———. 2010. Third National Poverty Assessment. Maputo.
Ndulu, B., and S. O’Connell. 2003. Revised Collins/
Bosworth Growth Accounting Decompositions. AERC
Growth Project, African Economic Research Consor-
tium, Nairobi.
Sonne-Schmidt, C., C. Arndt, and M. Magaua. 2008. Contri-
bution of Mega-Projects on GDP. Ministry of Planning
and Development, Direcção Nacional de Estudos e
Análise de Políticas (DNEAP), Maputo.
Sulename, Jose. 2001. “Economic Decline: A Study
with Reference to Mozambique.” Ph.D. diss., Department
of Economics, University of Notre Dame, South Bend, IN.
Tahari A., B. Akitoby, and E. Aka. 2004. “Sources of Growth
in Sub-Saharan Africa.” IMF Working Paper 04/176,
International Monetary Fund, Washington, DC.
UN COMTRADE (various years). United Nations Com-
modity Trade Statistics Database. http://comtrade
.un.org/db/
Vitek, Francis. 2010. “Economic Growth in Mozambique:
Experience and Policy Challenges.” International Mone-
tary Fund, Africa Department, Washington, DC.
World Bank. 2008a. Expanding the Possible in Sub-Saharan
Africa: How Tertiary Institutions Can Increase Growth and
Competitiveness. Poverty Reduction and Economic
Management (PREM), Washington, DC.
———. 2008b. Mozambique, Beating the Odds: Sustaining
Inclusion in a Growing Economy. A Mozambique Poverty,
Gender, and Social Assessment. Poverty Reduction and
Economic Management (PREM), Washington, DC.
———. 2010a. Mozambique Country Economic Memoran-
dum: Reshaping Growth and Creating Jobs through Trade
and Regional Integration. Washington, DC.
———. 2010b. Mozambique Investment Climate Assessment
2010. Washington, DC.
———. Various years. World Development Indicators.
Washington, DC.
79
 CHAPTER 4
Botswana’s Success: Good Governance,
Good Policies, and Good Luck
Michael Lewin
ver the past 60 years, Botswana’s economy has
O
been one of the most successful in the world. The
country’s achievement is remarkable, because at
independence, in 1966, its prospects were not encouraging.
ECONOMIC AND SOCIAL INDICATORS
Botswana is a sparsely populated, arid, landlocked country;
at independence it was also one of the poorest countries in
the world, with per capita income of just $70 a year. In the
first few years of independence, about 60 percent of current
government expenditure consisted of international devel-
opment assistance. There were only 12 kilometers of paved
roads, and agriculture (mostly cattle farming for beef pro-
duction) accounted for 40 percent of gross domestic
product (GDP).
By 2007 Botswana had 7,000 kilometers of paved roads,
and per capita income had risen to about $6,100 ($12,000
at purchasing power parity), making Botswana an upper-
middle-income country comparable to Chile or Argentina.
Its success is also evident in other measures of human
development. At independence, life expectancy at birth was
37 years. By 1990 it was 60, 10 years above the African aver-
age. Under-five mortality had fallen to about 45 per 1,000
live births in 1990, compared with 180 for Africa as a whole.
Development assistance has shrunk to less than 3 percent of
the government budget, and agriculture currently accounts
for only about 2.5 percent of GDP. Major strides have also
been made in infrastructure and education.1
Annual growth in per capita income averaged 7.0 percent
between 1966 and 1999 (table 4.1 and figure 4.1). The coun-
try’s performance is particularly impressive compared with
that of other African economies (figure 4.2).
Critics have argued that the social gains from this
growth have been somewhat limited. In fact, in addition to
the gains in health and life expectancy noted above, there
have been gains in poverty reduction and education. The
proportion of poor people fell from about 50 percent in
1985 to 33 percent in 1994, and the proportion of people
completing at least primary school rose from less than
2 percent at independence to about 35 percent in 1994
(Leith 2005).
Not all indicators are as positive: income distribution in
Botswana remains very unequal (the Gini coefficient was
about 0.55 in 1994).2 Unemployment remains high, reflect-
ing to a large extent rural to urban migration, although it
too has fallen, dropping from about 21 percent in the 1990s
to about 17 percent in 2008.
DIAMONDS AND DEVELOPMENT
Botswana’s extraordinary growth was fueled by minerals,
particularly diamonds. At independence, beef, the country’s
main export and largest sector, contributed 39 percent of
GDP. From independence until the 1970s, international aid
dominated the government budget and was the main source
of foreign exchange. At that time the mineral sector, mainly
diamonds, began to take off and soon became the dominant
81
 Figure 4.2 Average per Capita Income in Africa and
Table 4.1 Annual Growth in per Capita Income in
Botswana, 2001–09
Selected Economies, 1966–99
(purchasing power parity at current
Economy Average growth rate dollar prices)
Botswana 7.0
Chile 2.1
16,000
Hong Kong SAR, China 4.6
Indonesia 3.8 14,000
Ireland 4.1
12,000
Korea, Rep. of 6.1
Singapore 6.2 10,000

Dollars
Thailand 4.6 8,000
Source: Adapted from Leith 2005. 6,000
4,000

Figure 4.1 Average Annual Growth in per Capita Income 2,000

in Botswana, 1980–2008 0
2001 2002 2003 2004 2005 2006 2007 2008 2009
8 Africa Botswana
7 Source: OECD n.d.
6
5
Percent

4 “good economics, bad politics.” In almost all developing

3 countries, the government owns the mineral resources and
2
1
0
1980–90 1990–2000 2000–08
Source: Leith 2005.
sector. Income growth and the growth of the mining sector
accelerated in tandem from about 1974/75 until recently.
Luck—the discovery of diamonds as well as other
important minerals—was clearly an element of Botswana’s
success. However, there is considerable disagreement over
whether the discovery of minerals in a developing economy
generally brings good luck or bad (box 4.1).3 Indeed, many
countries in Africa, including Zambia, Nigeria, the Demo-
cratic Republic of Congo, Sierra Leone, and others, have
squandered vast amounts of their natural wealth. One does
not have to be persuaded that natural resources are neces-
sarily a curse to conclude that they certainly are not sufficient
for economic success or even a good predictor of it. In the
case of Botswana, minerals turned out to be lucky, but other
key ingredients in the recipe for success were present, includ-
ing good governance and good economic management.
Avoiding the bad governance curse
Mineral-based countries seem to be prone to bad govern-
ment, a phenomenon Acemoglu and others have termed
is therefore the main recipient of the revenues from their
extraction. This concentration of revenues with the govern-
ment as the conduit of benefits to the rest of the economy
can lead to a host of problems, including rent-seeking, cor-
ruption, and the efficiency losses that result from them. It is
not surprising that, with easily accessible wealth concen-
trated in the hands of the government, malignant dictator-
ships and predatory regimes often thrive. There also seems
to be a greater tendency for armed civil conflict in mineral-
based economies (Collier and Hoeffler 2004). These “bad
politics” are extremely costly.
Botswana has avoided these manmade disasters. Part of
its success may reflect luck: Botswana’s relatively homoge-
neous population has less potential for ethnic polariza-
tion, which when combined with mineral rents can be
particularly combustible. Still, most of the credit must be
given to the leadership, which, since independence, has
designed and fostered the conditions of governance that
have ensured stability and social and economic progress.
The government established respect for property rights
and the rule of law. It maintained a high degree of trans-
parency, which was reinforced by continuing the Tswana
tribal tradition of consultation. These consultative insti-
tutions, known as kgotla, created a degree of trust in the
government—the sense that government exists to serve
the people and promote development and is not the
instrument of one group or individuals for the purpose of
getting hold of the wealth. Tswana tradition also

82 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK
Box 4.1 The Resource Curse
The counterintuitive notion that the endowment of nat-
ural resources is a curse rather than a blessing has
become received wisdom. Sachs and Warner (1995)
report regression results showing that being a natural
resource or mineral exporter reduces a country’s devel-
opment prospects. Many other researchers find evidence
to support this claim (see for example, Auty 2004 and
the references therein). The fact that most of the East
Asian tigers (the Republic of Korea; Taiwan, China;
Hong Kong SAR, China; and Singapore) plus Japan are
resource poor has added to the perception that endowed
wealth is an obstacle to growth and development.
There is certainly evidence that a disproportionate
number of the poorest countries are dependent on
mineral exports. The direction of causality, however, is
far from clear: does mineral dependence retard devel-
opment, or does retarded development create mineral
dependence?
Accounting for the endogeneity of mineral depen-
dence overturns the Sachs-Warner conclusions. Coun-
tries that fail to develop remain resource dependent, but
resource abundance may be correlated with higher
growth (see, for example, Brunnschweiler and Bulte
2008; Alexeev and Conrad 2009). Many rich countries—
including Canada, Australia, Norway, the United States,
and Latin America’s top performer, Chile—were or are
mineral dependent. In Africa the two richest countries,
South Africa and Botswana, owe much of their wealth
to gold and diamonds.
In short, resources are not necessarily a curse, but
many countries have squandered their resources, and in
some cases, the presence of mineral resources may have
contributed to economic stagnation or decline. What is
it about mineral endowments that can make things go
wrong?
Dutch Disease
Dutch disease refers to the deleterious effects that pur-
portedly result from the real appreciation of the cur-
rency caused by a booming resource export sector
(Corden and Neary 1982). The discovery of minerals
or an increase in their international price leads to an
increase in income and expenditure. As long as some of
the increased expenditure goes to domestically pro-
duced goods, there will be a real appreciation of the
currency, which causes productive resources to move
from nonmineral traded goods, the output of which
declines, to nontraded goods, the output of which
increases. The resulting excess demand for tradables
CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK
can be financed by the revenue from the mineral
exports.
Because the entire process is initiated by an increase
in income, expenditure, and consumption, it is not clear
why it should be thought of as a disease. One reason
given is the presence of positive externalities in the
declining sector. The declining traded goods sector is
thought to have spillover effects that contribute to long-
run growth and industrialization. Hence, the real
appreciation caused by mineral exports is said to cause
“deindustrialization,” contributing to poor long-run
performance. Of course, in Africa and Botswana in
particular, the declining traded goods sector was agri-
culture, not industry (there was no significant industry
to deindustrialize). Some might argue that mineral
exports prevent the real depreciation needed to indus-
trialize. Additionally, it is argued, the mineral sector
fails to stimulate the nonmineral sector through link-
ages to the rest of the economy.
Against this effect, one must weigh the advantages
that minerals bring in terms of increased saving and
hence investment, which should offset any loss of com-
petitiveness caused by the real exchange rate. Moreover,
because government is usually the chief beneficiary of
the resource revenues, the increase usually eases con-
straints on public goods, such as infrastructure, educa-
tion, and health, which should contribute to growth
and industrialization. If private investment or public
goods expenditure fails to materialize, the problem lies
not in the real exchange rate but rather in the failure of
financial markets to allocate the saving and investment
productively or of government to invest wisely.
The Volatility Curse
Mineral prices tend to be volatile, which, it is argued,
destabilizes the economy (Hausmann and Rigobon
2003). This volatility affects the economy through the
exchange rate, which appreciates in the boom and
depreciates in the bust, and through fiscal policy, which
tends to be procyclical, leading to inefficient “stop-go”
provision of government services and infrastructure
projects.a It is argued that the perceived riskiness that
this volatility engenders leads to lower investment, par-
ticularly foreign direct investment, which seeks out
safer markets. In addition, governments tend not only
to increase spending when current revenues rise but often
also to believe that revenues will continue to rise; they
therefore borrow against future revenues. Instead of
being able to use accumulated assets to help in the bust,
(continued next page)
83
 Box 4.1 (continued)

economies find themselves contending with debt Exhaustibility

overhang.
It is possible to insulate the domestic market from
the volatility of the commodity markets. Because the
government is the conduit of the cycle to the economy,
it is also the solution. The key is to delink expenditure
and revenue, essentially by saving all or part of the rev-
enues, thereby replacing a volatile income stream with a
more stable one. Expenditure can then be based on this
more stable, permanent income. In this way, consump-
tion can be smoothed over time, in some cases over
generations. Many countries, including Botswana,
Chile, Kuwait, Norway, Qatar, and Saudi Arabia, have
successfully applied such an expenditure-smoothing
policy.
The Governance Curse
In most developing countries the government claims
ownership of all minerals. If government institutions
are weak to begin with, there is a strong likelihood that
the resources will be squandered. In addition, the pres-
ence of the resources may weaken and corrupt govern-
ment. The relative ease with which mineral revenues are
collected can lead to a lack of transparency and
accountability. The resulting interdependence of gov-
ernment and the mining industry reinforces this ten-
dency. Government itself is seen as the means to
acquiring wealth, which typically leads to rent seeking
and corruption. There is also evidence that in many
cases mineral wealth provokes or fuels internal conflicts
(Collier and Hoeffler 2004).
In an ironic metaphor, underground minerals are
often referred to as one of the “commanding heights” of
the economy. Postindependence governments claimed
the need to control these commanding heights for the
benefit of all the governed. One of the unforeseen con-
sequences of this policy was the perception that control
of government is the path to wealth, which in turn,
unsurprisingly, has often led to regional and ethnic con-
flict for control of these resource rents (Collier and
Hoeffler 2004).
Minerals are a windfall, even in the long run, in the
sense that they are finite and that when they are
depleted an economic slump will occur. That a blessing
does not last forever hardly makes it a curse, however. If
at least some of the windfall is saved, in the form of
physical and human capital or financial assets, future
generations will benefit from it. Moreover, the deple-
tion usually does not occur suddenly. With some fore-
sight, the adjustment and transition can be cushioned
by accumulated saving.b
How Has Botswana Fared?
To a large extent, Botswana overcame the threat of the
resource curse:
1. Dutch disease was avoided, most significantly by
government investment in public goods and infra-
structure. The government also took measures to
help boost productivity, by limiting parastatals and
avoiding import substitution policies.
2. The volatility curse was overcome by unlinking pub-
lic expenditure from revenue. The government
established savings funds and avoided typical pro-
cyclical behavior and real exchange rate volatility.
3. The governance curse was avoided, through a series
of actions described in the chapter. Botswana’s suc-
cess in this regard is exceptional.
4. The question of whether the exhaustibility chal-
lenge has been met remains unclear. If current pre-
dictions are correct, diamond revenues will begin
declining in 2016 and could be exhausted by 2029.
Botswana has adopted many of the appropriate
policies to prepare for the depletion of its mineral
base, by accumulating funds for the future, building
infrastructure, and investing in health and educa-
tion. However, the growth of the nonmineral traded
goods sector, which has remained flat at about
5 percent of GDP, will need to accelerate to avoid
balance of payments problems as diamond revenues
begin to decline.

a. Rather than being procyclical, in most developing countries, the government initiates the cycle because it is the recipient of the
rents. The boom therefore begins with government expansion and the bust with its contraction. Hence, real exchange rate vari-
ability is also the result of volatile government expenditure.
b. A full discussion of mineral resource management in the context of intertemporal and intergenerational allocations goes
beyond the objectives of this chapter. Essentially, the optimal management of resource rents requires expenditure smoothing
not only to avoid cyclical volatility but also to spread the benefits of the windfall over time. Like a rational household, an opti-
mizing government will attempt to base its long-run expenditure on some calculation or estimate of its “permanent” or stable
revenue rather than on its current or transitory revenue. For some calculations of permanent income in the case of Botswana,
see Basdevant (2008).

84 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK
respected private property; the fact that many of the tribal
leaders who helped usher in modern government were
also large cattle owners may have reinforced this respect.
Acemoglu and Robinson (1999) emphasize that property
rights and the rule of law are the key factors explaining
development success. They also emphasize the importance
of the preindependence colonial regime in determining the
adoption of these institutions after independence. More
heavily settled colonies tended to establish institutions to
maintain the rule of law and enforce property rights; in
sparsely settled colonies, the colonial regimes tended to be
more “extractive” or “exploitative,” unconstrained by law
and respect for property. Although Botswana’s colonial past
is not typical of the kind of regime that usually led to a
postindependence government constrained by law and
property rights, its unique colonial history had the same
effect.4 Before independence, Botswana was a British pro-
tectorate, the colonial power having been “invited” in.
Because Botswana was not colonized for economic or
strategic advantage, the colonial rulers, it is argued, did not
impose the extractive-type regime often found in other
sparsely populated areas. Thus the regime that evolved after
independence was one that respected the law and property
and was dedicated to development.
More controversial is the role of democracy. Botswana
has maintained a parliamentary democracy since indepen-
dence. To be sure, its democracy is not perfect: the country
has had one-party rule since independence, women’s repre-
sentation is limited, and there has been some criticism that
minorities (particularly the San people) are not treated
equally. Nevertheless, the government functions in a demo-
cratic manner, elections are “free and fair,” and the govern-
ment is responsive to the electorate and transparent in its
dealings.
Is democracy a positive factor for economic growth and
development? None of the miracle East Asian economies or
Latin America’s high performer (Chile) was democratic
during the first years of rapid growth and development,
leading many commentators to conclude that property
rights and the rule of law are more important than democ-
racy (Acemoglu, Johnson, and Robinson 2003). For some
observers, democracy is actually a hindrance to economic
development. Although this conclusion is probably unwar-
ranted, the fact that many democracies (such as India) have
not fared well economically undoubtedly means that
democracy is neither sufficient nor necessary for growth
and development.5
In the case of mineral-rich countries like Botswana,
democracy may be an important catalyst. Mineral-dependent
CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK
economies seem to spawn the predatory type of dictatorship
rather than the relatively benevolent ones of East Asia’s past.
Although this notion is largely speculative, it seems that a
democratic government dependent on mineral wealth is
more likely than a dictatorial one to be responsive to devel-
opment needs, to settle disputes peacefully, and to respect
the rule of law. Very few, if any, mineral-rich countries in
Africa have been ruled as peacefully and productively as
Botswana; it is hard to escape the conclusion that the dem-
ocratic institutions established at independence are an
important part of the explanation.
Social science cannot rigorously assess the relative
importance or contribution of leadership in the evolution
of successful institutions. However, in the case of Botswana,
leadership, particularly that of its first president, Seretse
Khama, may have been crucial, especially in the areas of
mineral exploitation and the rights of the state versus those
of the tribes. The discovery of minerals can easily lead to
civil war and the dissolution of the state. To prevent this
from happening, even before independence, Khama’s party,
the Botswana Democratic Party (BDP), wrote into its plat-
form its intention to assert the state’s rights to all mineral
resources. After independence, the government reached
agreement on ownership of mineral resources with the
tribal authorities. Although the largest diamond deposits
were discovered in Khama’s own district of Bamangwato, he
chose the country over his tribal land, thus helping limit the
possibility of conflict.
Implementing good policies
All mineral-based economies face the issue of an appreciat-
ing real exchange rate (box 4.2). The well-known possible
harmful consequences of this are often referred to as Dutch
disease (see box 4.1). Prevention of Dutch disease in
Botswana consisted of three key components: fiscal saving, a
surplus on the current account of the balance of payments,
and heavy government investment in infrastructure and
human capital. Together these policies limited the erosion of
domestic productivity and competitiveness that can result
from the appreciation of the real exchange rate. High fiscal
saving limits current consumption, reducing pressure on
domestic price inflation, a typical problem in natural
resource booms. Some of the government saving took the
form of offshore investments, which directly limited real
exchange rate depreciation and diversified the sources of
future foreign exchange revenues. The accumulation of
reserves is a form of self-insurance against short-run
declines in mineral revenues as well as long-run reductions
85
 Box 4.2 Windfall Economics 101
The mechanism by which mineral resource flows affect
the real exchange rate depends on the nominal
exchange rate regime maintained by the country.
Botswana has a fixed exchange rate system. Initially, it
pegged its currency to the U.S. dollar. When the South
African rand depreciated relative to the dollar,
Botswana suffered from the appreciation of the pula
(Botswana’s currency) relative to the rand (the currency
of its main trading partner).a The pula is therefore now
pegged to the rand.
In a fixed-rate regime, the first effect of a resource
windfall is the creation of an excess supply of foreign
exchange. To maintain the peg, the monetary authority
must accumulate reserves. Doing so will inject domestic
a. In contrast, Argentina maintained a dollar peg after the devaluation of the Brazilian real. Many analysts consider this a key ele-
ment leading to the Argentine crisis.
in these revenues as a result of mineral depletion. The instru-
ments of public saving took the form of the Public Service
Debt Management Fund and the Revenue Stabilization
Fund. These funds have proved to be successful mechanisms
for managing fiscal saving; income from the Revenue
Stabilization Fund is now a stable source of government
revenue.
Another part of the fiscal saving was channeled to
domestic assets, combating the effect of the loss of compet-
itiveness by raising productivity. When this investment is
focused on public goods (for example, infrastructure,
health, and human capital), it will contribute to growth
without crowding out private sector investment and devel-
opment. Public sector saving was positive in every year
between 1975 and 1996, fluctuating between 10 and 40 per-
cent of GDP. Public sector investment was fairly constant at
about 10 percent of GDP.6 However, if one counts expendi-
ture on health and human capital as investment, government
investment has consistently remained about 20 percent of
GDP. The capital budget focused on basic infrastructure
with about 30 percent devoted to water, electricity, roads,
communication, and transportation. Twenty percent of the
capital budget, on average, went to education and around
30 percent of recurrent expenditures was devoted to educa-
tion and health.7 While quantifying the effectiveness of
public expenditure on growth and development is notori-
ously difficult, particularly in resource-based economies
where soft budget constraints may lead to overinvestment,
the emphasis on infrastructure, health, and education
86 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK
money into the economy, however, resulting in domestic
price inflation, which will eventually erode local compet-
itiveness until the balance of payments surplus is elimi-
nated. This is the effect of real appreciation when the
nominal rate is fixed. To prevent this from happening,
the government has to sterilize the effects of the resource
rents on the domestic money supply. In the medium to
long run, the fiscal authority can do so by absorbing the
excess domestic money through fiscal budget surpluses.
This policy of twin surpluses—a balance of payments
surplus in tandem with a fiscal budget surplus—is essen-
tially the policy that Botswana has followed. (Some ana-
lysts have argued that China has followed a similar policy
to maintain an undervalued currency.)
seems to have served Botswana well. As noted above, the
number of paved roads increased from around 20 kilome-
ters in 1970 to 2,300 in 1990; 90 percent of the population
had access to safe water (compared with 29 percent in
1970); and the number of telephone connections rose from
around 5,000 in 1970 to 136,000 in 2001.8 Similarly, in edu-
cational achievement Botswana leads Africa. For example,
adult literacy, male and female, is around 80 percent
compared with 69 percent and 50 percent, respectively, for
the rest of sub-Saharan Africa.9
As a result of its prudent fiscal policy, Botswana has
enjoyed relative macroeconomic stability and avoided the
boom-slump cycles that characterize many mineral-based
economies. Monetary policy was also restrained: for most of
the postindependence period, inflation was moderate, aver-
aging about 10 percent (Maipose n.d.). The periodic slow-
downs in the diamond industry have thus by and large not
been passed on to the rest of the economy. By withholding
some of the benefits from the economy during the booms,
the government has, to some extent, been able to insulate it
from the busts.10
Shunning import substitution and parastatals
Two things Botswana did not do are also significant. Unlike
many African countries, it did not adopt a policy of import
substitution, and it did not expand the extent of state-
owned productive entities, which employ only about 5 per-
cent of the workforce in Botswana (figure 4.3).
Figure 4.3 Percentage of GDP Accounted for by Botswana’s combination of policy and governance evidently
State-Owned Enterprises in Selected helped it avoid the worst effects of the resource curse.
African Countries, Early 1990s

30 PLANNING FOR A FUTURE

WITHOUT MINERALS
25
Botswana is still a mineral-based economy: it has not suc-
Percentage of GDP

20 ceeded in significantly diversifying its economy away from

diamonds. Government remains the largest employer
15
(employing 30 percent of the active workforce); other than
10 minerals (which employ a relatively small share of workers),
most production is in nontraded goods. This pattern is typ-
5 ical of mineral-dependent economies. Real GDP and mining
income have moved in tandem since 2000 (figure 4.4), and
0
mining consistently accounts for about 40 percent of GDP,
a

na

ia

go

e
an

ny

bi

bw
oo

an

completely dwarfing the contributions of manufacturing

ha

To

m
w

Ke

nz
er

ba
G

Za
ts

am

Ta

m
Bo

and agriculture (figure 4.5).11 Minerals, mostly diamonds,

Zi
C

Source: Leith 2005.

Figure 4.4 Mining Revenues and Real GDP in Botswana,
2000–09
Being part of the South African Customs Union (which 30,000
required cooperation with the then odious regime in South
Pula (millions) at constant

25,000
Africa) meant maintaining a fairly low tariff regime and
1993/94 prices

20,000
provided a steady stream of government revenue. Avoiding
an activist import-substituting policy and maintaining lim- 15,000
ited government involvement in production seems to have 10,000
paid off for Botswana by allowing it to avoid many of the
5,000
inefficiencies and structural deficits that so often arise from
0
such policies.
00

01

02

03

04

05

06

07

08

09
Trade policy is also an area where good governance and
20

20

20

20

20

20

20

20

20

20
good policies reinforce one another. A government rich with Mining Real GDP
mineral revenues is an inviting target for rent seekers and
worse; restricting the avenues for rent seeking and corrup- Source: Bank of Botswana 2010.

tion thus helps preserve the efficiency and integrity of the

government. Even if the theoretical merits of import substi- Figure 4.5 Contributions of Mining, Manufacturing, and
tution or the existence of state-owned enterprises seem per- Agriculture to Botswana’s GDP, 2000–09
suasive, in practice both often result in inefficiency and
45
drain fiscal resources.
40
Good fiscal policies by themselves may not be sufficient
35
for success. Many mineral-based economies with high rates 30
of investment have not enjoyed the positive results that
Percent

25
Botswana has. The quality of investment is evidently as 20
important as the quantity. Moreover, in the hands of the 15
venal and corrupt, government savings funds can easily turn 10
into slush funds for the favored elites. There may be a lesson 5
in this from Botswana: good governance will aid the effec- 0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
tiveness of good policy, and good policy encourages better
Agriculture Mining Manufacturing
government. Policy formulators should therefore not ignore
the political economy consequences of economic policy. Source: Bank of Botswana 2010.

CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK 87
 comprise about 85 percent of exports, and mineral revenues
account for about 50 percent of government revenues (Bank
of Botswana 2008).12
The diamond industry is expected to decline in the near
future, with revenues projected to begin falling in 2016 and
to be depleted by 2029 (Basdevant 2008). If these projections
are correct, adjustment is inevitable and unlikely to be pain-
less, although Botswana’s prudent policies—particularly
regarding saving and investment—have left it in a relatively
strong position to facilitate a soft landing. For most years,
domestic saving has been above 40 percent and investment
about 35 percent of GDP. (The difference is the surplus on
the current account of the balance of payments.) This
implies that from the perspective of the economy as a whole,
all revenue from minerals is being saved rather than con-
sumed, because total saving from national income is roughly
equal to mineral income. When mineral revenues begin to
decline, some of the adjustment could thus be absorbed by
saving rather than consumption. Thus, even if growth slows,
consumption growth can be maintained as the proportion of
consumption to income rises.
The government has also been a big saver. Central bank
profits and other income from accumulated assets now
make up about 30 percent of government revenue, most of
it income from the accumulated saving of mineral revenues.
At times, the overall fiscal surplus rose to more than 30 per-
cent of GDP (Maipose n.d.), and public sector investment
(as noted) has been consistently about 10 percent of GDP. If
investment is defined more broadly to include human cap-
ital and health, public investment rises to about 20 percent
of GDP, which meets the IMF benchmark for sustainability
(Basdevant 2008). The government has used this accumu-
lated savings to smooth expenditure over the business
cycle, thus providing the economy with a short-run shock
absorber. However, as argued in box 4.1, the accumulated
saving is also a long-run shock absorber, because it has
built up human and physical capital and financial assets
that should serve the economy well during the adjustment
period. In addition, the accumulated foreign assets man-
aged by the Bank of Botswana now amount to about
40 months of imports.
These advantages notwithstanding, Botswana faces seri-
ous challenges. Dependence on mineral exports is the key
weakness of the current economic outlook and diversifica-
tion of the traded goods sector is the most important policy
objective. Accumulated foreign assets and the income from
them can ease the welfare cost of adjustment, but it cannot
eliminate it. Barring a miraculous jump in productivity,
some real depreciation of the pula and the consequent hard-
88 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK
ships this may cause, will be necessary to create the compet-
itiveness needed to replace falling mineral exports.
In 2009 manufacturing contributed less than 5 percent
of GDP (see figure 4.5). This share will have to rise if
Botswana is to avoid decline. The fact that manufacturing
has been flat as a proportion of income conceals its growth,
because just keeping pace with Botswana’s growing econ-
omy has required rapid growth.13 However, manufacturing
growth will have to accelerate if Botswana is to raise or even
maintain its standard of living in the long run.
Key to such growth will be Botswana’s ability to attract
foreign direct investment (FDI). Although FDI fell between
2002 and 2007 (Bank of Botswana 2010), Botswana’s repu-
tation as a stable country and successful economy should
help it attract FDI in the future.14 Policy makers should
examine the country’s overall investment environment with
a view to attracting desirable investment.15
Clouding Botswana’s smooth transition to a more diver-
sified economy is HIV/AIDS, which hit Botswana harder
than any country in Africa (with the possible exception of
South Africa). The epidemic has reversed many of the
impressive gains in health indicators achieved over the past
100 years and reduced productivity. The number of deaths
from AIDS began declining in about 2003 (World Bank
2009), but the costs and consequences of the epidemic will
persist for many years.
CONCLUSION
Landlocked Botswana seems to have defied the odds by
creating a successful economy. Poverty has been reduced,
education has become more widespread, and health indi-
cators had improved before the HIV/AIDS epidemic undid
some of that progress.
The country’s vast natural resources played a key role in
this accomplishment, but the mere endowment of resources
is clearly not the whole story. In much of Africa—and in
other parts of the world—natural resources have not always
been conducive to growth and development; in many cases
they seem to have brought out the worst in countries, in the
form of conflict and predatory governments.
Studying the effect of mineral wealth on economic out-
comes is timely, because the increase in the prices of natu-
ral resources as a result of the rise of China and India is
likely to result in windfalls for many African countries.
How can countries turn these windfalls into long-run
growth and development? In Botswana’s case, the key to
successfully harnessing natural resources lay in good gover-
nance and good policies. Governance has not been perfect
in Botswana, but it has been good. Botswana has been
largely free of kleptocracy and civil conflict; it has main-
tained a transparent, law-abiding government; and it has
implemented good policies, including a hyper-prudent fis-
cal policy, which has done much to diversify foreign
exchange earnings and prevent the volatility that typifies
many resource-based economies. Investments in human
and physical capital and vast improvements in infrastruc-
ture have also raised Botswana’s productivity, which,
together with its substantial financial reserve in the form of
foreign assets, should help ease the transition to a more
diversified economy.
NOTES
1. For discussions of Botswana’s success, see Maipose
(n.d.), Acemoglu and Johnson (2003), and Leith (2005).
2. Leith (2005) notes that when measured accurately, tak-
ing into account social, health, and educational services pro-
vided by the government to the poor, this measure falls to
about 0.53. No measure of inequality is without serious ana-
lytical problems; the Gini coefficient is probably the best
available and most widely used index. The coefficient ranges
from 0 (perfect equality) to 1 (perfect inequality). The
higher the index, the more unequal the society. An index
above 0.5 is thought to denote an unequal distribution
(Leith 2005).
3. See, for example, Auty (2004) and Sachs and Warner
(1995). For a contrary view, see Stevens (2003) and
Brunnschweiler and Bulte (2008). See also Lederman and
Maloney (2007).
4. The analysis is based on the idea that some colonial
regimes were mostly “extractive”—that is, the regime
existed to reap the maximum out of the colony’s economy.
Other regimes, usually ones in which there were more colo-
nial settlers, were less extractive. They often established
institutions that put more constraints on the extractive and
arbitrary powers of government. These constraining institu-
tions often carried over into independence.
5. All of the high-performing developing countries men-
tioned have established strong democratic institutions. The
question therefore arises whether the causality runs the
other way, from growth to democracy. It can also be argued
that democracy is necessary to sustain a high-income econ-
omy (see Barro 1997).
6. Isham and Kaufmann (1999) estimate this proportion
(10 percent of GDP) to be the likely ceiling for public invest-
ment to remain productive. Beyond this they find public
sector investment is likely to be detrimental to growth.
7. Lange and Wright (2002) point out that the only single
“nonproductive” item of comparable weight in the budget
CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK
was defense expenditure, which average about 11 percent of
total expenditure. While this is high relative to most coun-
tries Botswana arguably had good reason to do this.
8. World Bank cited in Lange and Wright, 2002, 32.
9. World Bank 2009. Botswana, however, lags other upper
middle income countries in these categories.
10. Botswana has not been able to escape the effects of the
current global crisis: GDP declined 4 percent in 2009, and
for the first time in many years, Botswana had fiscal and
balance of payments deficits. However, the economy
appears to have weathered the worst, with the Bank of
Botswana projecting real growth of more than 3 percent for
2010 and 2011.
11. Mining’s share fell in 2009 because of the slump in
world trade, not because of the declining importance of the
sector.
12. Between 2005 and 2008, before the effects of the global
slump set in, exports grew by about 40 percent. The share
of diamonds fell from about 75 percent in 2005 to about
65 percent in 2008. Total mining remained more or less
constant, however, at about 85 percent of total exports. The
share of diamonds fell because of rising copper and nickel
exports (Botswana Central Statistical Office 2009).
13. Other areas of interest are services, including finan-
cial services; downstream diamond processing and trad-
ing; and tourism, which represented about 10 percent of
GDP in 2008 and is a potential growth sector for export
earnings.
14. FDI rose again in 2008. It is too early to tell whether this
is a trend.
15. Botswana had the best ranking in Africa on Trans-
parency International’s corruption perception index in 2011
and ranked 33rd worldwide. (See http://www.transparency
.org/policy_research/surveys_indices/cpi/2010.) On the
World Bank’s 2010 Ease of Doing Business index, Botswana
ranked 3rd regionally but only 52nd worldwide, indicating
some room for improvement (see http://www.doingbusiness
.org/rankings).
REFERENCES
Acemoglu, D., S. Johnson, and J. A. Robinson. 2003. “An
African Success: Botswana.” In In Search of Prosperity:
Analytical Narratives on Economic Growth, ed. D. Rodrik.
Princeton, NJ: Princeton University Press.
Acemoglu, D., and J. A. Robinson. 1999. “On the Political
Economy of Institutions and Development.” American
Economic Review 91 (4): 938–63.
Alexeev, Michael, and Robert Conrad. 2009. “The Elusive
Curse of Oil.” Review of Economics and Statistics 91 (3):
586–98.
89
 Auty, R. (ed.) 2001. “Resource Abundance and Economic
Development.” Wider Studies in Development. Oxford
University Press, Oxford, UK.
Bank of Botswana. 2008. Annual Report. Gabarone.
——. 2010. Botswana Financial Statistics. Gaborone. http://
www.bankofbotswana.bw/index.php/content/2009
110615035-botswana-financial-statistics.
Barro, R. 1997. Determinants of Economic Growth: A Cross-
Country Empirical Study. Cambridge, MA: MIT Press.
Basdevant, O. 2008. “Are Diamonds Forever? Using the Per-
manent Income Hypothesis to Analyze Botswana’s
Reliance on Diamond Revenue.” IMF Working Paper,
International Monetary Fund, Washington, DC.
Botswana Central Statistical Office. 2009. External Monthly
Trade Digest January. Gaborone. http://www.cso.gov.bw/.
Brunnschweiler, C. N. and E. H. Bulte. 2008. “The Resource
Curse Revisited and Revised: A Tale of Paradoxes and
Red Herrings.” Journal of Environmental Economics and
Management 55: 248–64.
Collier, P., and A. Hoeffler. 2004. “Greed and Grievance in
Civil War.” Oxford Economic Papers 56, Oxford Univer-
sity, Oxford, UK.
———. 2005. “Democracy and Resource Rents.” Oxford
University, Oxford, UK.
Corden, W. M., and P. J. Neary. 1982. “Booming Sector and
Deindustrialization in a Small Open Economy.” Eco-
nomic Journal 92: 825–48.
Hausmann, R., and R. Rigobon. 2003. “An Alternative Inter-
pretation of the ‘Resource Curse’: Theory and Policy
90 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK
Implications.” NBER Working Paper 9424, National
Bureau of Economic Research, Cambridge, MA.
Isham, J., and D. Kaufmann. 1999. “The Forgotten Rationale
of Policy Reform: The Productivity of Investment Proj-
ects.” Quarterly Journal of Economics 116 (1).
Lange, G., and M. Wright. 2002. “Sustainable Development
in Mineral Economies: The Example of Botswana.”
Ceepa Discussion Paper Series, University of Pretoria,
Pretoria, South Africa.
Lederman, D., and W. F. Maloney, eds. 2007. Natural
Resources: Neither Curse nor Destiny. Washington, DC:
World Bank.
Leith, J. C. 2005. Why Botswana Prospered. Montreal:
McGill-Queen’s University Press.
Maipose, G. n.d. Policy and Institutional Dynamics in Sus-
tained Growth in Botswana. University of Botswana,
Gaborone.
OECD (Organisation for Economic Co-operation and Devel-
opment). n.d. Africa Economic Outlook. http:// www
.africaneconomicoutlook.org/en/countries/southern-
africa/botswana/.
Sachs, Jeffrey D., and Andrew M. Warner. 1995. “Natural
Resource Abundance and Economic Growth.” NBER
Working Paper W5392, National Bureau of Economic
Research, Cambridge, MA.
Stevens, P. 2003. “Resource Impact: Curse or Blessing? A Lit-
erature Survey.” Journal of Energy Literature 11 (1): 3–42.
World Bank. 2009. Country Partnership Strategy for the
Republic of Botswana. Washington, DC.
 CHAPTER 5
Mauritius: An Economic Success Story
Ali Zafar
n 1961 James Meade, a Nobel Prize recipient in econom-
I ics, famously predicted a dismal future for Mauritius
because of its vulnerabilities to weather and price shocks
and lack of job opportunities outside the sugar sector. The
small island nation in the Indian Ocean of approximately
1.3 million people has defied those predictions, trans-
forming itself from a poor sugar economy into a country
with one of the highest per capita incomes among African
countries.1 Mauritius’s combination of political stability,
strong institutional framework, low level of corruption,
and favorable regulatory environment has helped lay the
foundation for economic growth, while its open trade
policies have been key in sustaining growth. The govern-
ment functions as a parliamentary democracy, and the
country has an efficient administration that is both tech-
nically competent and adaptive to changing global
economic circumstances. Mauritius’s financial sector is
sufficiently well developed that it has begun to position
itself as a platform for investment linking East Africa with
India and China.
Headline figures related to Mauritius’s economic perfor-
mance are impressive. Growth of real gross domestic prod-
uct (GDP) has averaged more than 5 percent a year since
1970, and annual growth in real per capita income has like-
wise been strong. GDP per capita increased about sevenfold
between 1976 and 2008, from less than $1,000 to roughly
Ali Zafar is a macroeconomist in the World Bank Group.
$7,000. Imports and exports have boomed; together, they
reached more than 100 percent of GDP during the late
1990s and early 2000s. At the same time, efforts at eco-
nomic diversification have been successful, allowing the
country to move from sugar to textiles to a broader service
economy. Mauritius’s reliance on trade-led development
has helped the country achieve respectable levels of export
performance. Along the way, measures of human develop-
ment have improved substantially.
Despite being a small island economy vulnerable to
exogenous shocks, Mauritius has been able to craft a strong
growth-oriented developmental path. Natural disasters and
terms-of-trade shocks have never prevented the economy
from having strong and regular growth. Constrained at
inception by a monocrop sugar economy, low amounts of
arable land, and a high rate of population growth, Mauritius
has emerged as a regional entrepôt and tourism destination
as well as the top-ranked African country in the World
Bank’s Doing Business survey (in 2010, it ranked 20th of
183 countries).
Although a variety of explanations have been advanced
to explain Mauritius’s growth, the country’s focus on inter-
national trade has been, without doubt, a critical element of
that performance.2 Mauritius’s preferential access to trading
partners, particularly the European Union (EU), in the
sugar, textile, and clothing sectors resulted in subsidies for
91
 the export sector and provided important foreign exchange
for the economy. Preferential trading deals accounted for
strong growth in Mauritius’s total exports between the
1970s and the 1990s. Although imports tariffs in Mauritius
were kept high, they have never been high enough to inter-
fere with the overall trade regime. For example, the average
tariff for manufacturing was 86.2 percent in 1980, but it
fell to 30.1 percent in 1994 (Wignaraja and O’Neil 1999).
Simultaneously, Mauritius pursued a very liberal invest-
ment regime and used incentives to attract foreign direct
investment (FDI). Mauritius also offset the burdens on its
exporters with tariff-free access for productive inputs, with
tax incentive subsidies and relaxed labor market regula-
tions in the export sector. Furthermore, it has used export
processing zones (EPZs) to export key manufacturing
goods, mostly apparel and textiles. Mauritius’s overall trade
and investment policy has been based on a managed embrace
of globalization and cultivation of market access.
Aside from its trade and investment policy, Mauritius has
benefited from prudent fiscal, exchange rate, and monetary
policy, the latter of which has also been beneficial to export
performance. To compensate for the myriad disadvantages
of limited scale, Mauritius has developed a plethora of
strong institutions and good governance. Additionally, the
public and private sectors maintain a vibrant partnership
that manifests itself in a range of areas.
Finally, Mauritius has always displayed receptivity to
new ideas and adaptability. At various points in its history,
Mauritius has used intervention, subsidies, and targeting
to adapt to shifting economic circumstances. Although
Mauritius benefited from the sugar protocols, the
government also recognized early on the advantages of
diversification. As a result, it relied heavily on EPZs but
ensured that there was no anti-export bias. Mauritius has
also proven very adept at embracing new sectors, particularly
light manufacturing, offshore banking and financial services,
and service-related information and communication tech-
nology (ICT). It has adapted and transformed its ethnic
pluralism into a tangible economic force.3 It has forged a
vibrant democracy with competing parties, a strong electoral
system, and an open media.4 In sum, Mauritius’s impressive
economic performance has not been an accident, but rather
the result of careful planning and policies. During and in the
aftermath of the biggest exogenous shocks to its economy in
recent times—the phasing out of the Multifibre Arrange-
ment governing textiles, significant reductions in EU sugar
protocol prices, the 2008 food and fuel crisis, and the
2008–09 global financial crisis—Mauritius’s economy has
displayed strong resilience.
THE GROWTH STORY
Since the 1970s Mauritius has recorded very high growth
rates and sustained increases in human development
indicators through a combination of good macroeco-
nomic policies and strong institutions (table 5.1). With
the advent of the sugar preferences and the EPZs in the
1970s and 1980s, Mauritian authorities have succeeded in

Table 5.1 Major Economic Indicators for Mauritius 2006–09

Indicator 2006 2007 2008 2009
National income (%)
Real GDP growth 3.9 5.4 4.2 1.5
Consumer prices 11.8 8.7 6.8 4.0
Unemployment 9.1 8.5 7.2 8.3
Fiscal (% of GDP)
Revenue 19.3 22.0 22.5 21.1
Expenditure 23.6 25.7 26.1 25.6
Fiscal balance –4.6 –4.0 –3.4 –4.5
Total external debt 6.1 4.6 6.1 8.0
Total domestic debt 46.6 45.6 43.8 42.0
External (%)
Import 16.4 6.0 20.6 –24.5
Export 8.9 –4.7 7.3 –23.5
Current account (% of GDP) –9.4 –5.6 –10.4 –8.1
Terms of trade –5.8 –1.0 –1.1 1.6
Domestic investment (% of GDP)
Public 6.9 6.4 5 6.1
Private 17.3 18.7 19.6 20.2
Nominal GDP ($ billions) 6,507 7,521 9,321 8,527

Source: IMF (2010); Government of Mauritius.

92 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY

transforming the economy and laying the foundation for
stable growth in the future. Sugar and textile revenues
have been used to facilitate service-sector development
and contribute to socioeconomic progress and higher liv-
ing standards.
A well-managed economic regime
Proper macroeconomic management—fiscal discipline
during boom times, monetary management that kept infla-
tion in the single digits and that produced interest rates
that encouraged domestic savings, and an exchange-rate
policy that maintained flexibility and competitiveness for
exporters—have all been key to Mauritius’s economic suc-
cess. At the same time, flexible public policy, especially in
the form of experiments creating EPZs in the 1980s and
embracing the ICT industry in the 2000s, was also an
important feature. Trade reforms, the development of a
social welfare system, and policies that favored human cap-
ital development also played a part. The result has been
manageable fiscal and current account deficits, high rates of
private investment, and respectable and stable growth rates.
A very successful economic trajectory
Between 1977 and 2009, and despite the volatility caused by
exogenous shocks during those years, real GDP in Mauritius
grew on average by 5.1 percent annually, compared with
3.2 percent for Sub-Saharan Africa (figure 5.1). The growth
rate accelerated in the 1980s as a result of the macroeco-
nomic reforms instituted in response to protracted balance
of payments and fiscal troubles. Following the reforms,
Figure 5.1 Real GDP Growth in Mauritius, 1981–2009
12
10
8
Percent
0
81
83
85
87
89
91
93
19
19
19
19
19
19
19
19
Source: Government of Mauritius.
Mauritius experienced steady growth, low inflation, and
increased employment. GDP per capita, meanwhile,
increased approximately sevenfold between 1976 and 2008,
from less than $1,000 to nearly $7,000 (figure 5.2). At the
same time, consumer price inflation in Mauritius has
remained in the low single digits through the 1990s and
early 2000s, and the country’s debt stock has been manage-
able.5 Windfalls from sugar and textile preferences have
been used wisely to help promote diversification and boost
growth. The structural transition away from agriculture and
into manufacturing and services shows the success of the
government’s efforts at economic diversification.
Favorable international comparisons
Both a cursory examination of economic indicators and
detailed diagnostics show that Mauritius has been one of
the best-performing and most stable economies in Sub-
Saharan Africa in recent decades, even in the presence of
terms-of-trade and oil shocks. Using purchasing power
parity (PPP) data for 44 Sub-Saharan African countries,
Arbache and Page (2008) examine country-level dynamics
of long-run growth between 1975 and 2005 and conclude
that Mauritius was one of the best performers, both in per
capita growth performance and in low volatility of growth,
alongside Botswana, Cape Verde, Gabon, Namibia, the
Seychelles, and South Africa.
Decomposing the standard deviation of GDP per capita
and economic growth into within-country and between-
country components, one finds that Mauritius’s growth
stability was much higher than that of many oil economies,
such as Angola and Nigeria, and of many resource-poor
95
97
99
01
03
05
07
09
19
19
20
20
20
20
20
CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 93
 Figure 5.2 GDP per Capita in Mauritius, 1976–2008 than in the past and to a greater extent by productivity
growth, with TFP growth during this period averaging
8,000 more than 1 percent a year.6 Their estimates suggest that in
7,000 the EPZ sector, TFP grew by 5.4 percent from 1991 to 1999.
6,000 In another study, Rojid and Seetanah (2009) provide
evidence that TFP gains in Mauritius have reached a
U.S. dollars

5,000
4,000 plateau. TFP growth averaged 1.4 percent in the 1980s,
3,000
1.0 percent in the 1990s, and 0.7 percent in the 2000s, but
the TFP contribution to overall growth varies depending
2,000
on the different growth rates (table 5.2). For the entire
1,000
period 1980–2007, TFP contribution to growth in Mauritius
0
averaged 1.0 percent annually, considerably higher than the
76

80

84

88

92

96

00

04

08
0.3 percent average in Common Market for Eastern and
19

19

19

19

19

19

20

20

20
Source: Government of Mauritius.
economies, such as Burundi and Central African Republic,
in Sub-Saharan Africa. While Mauritius was growing,
much of Sub-Saharan Africa was not. In 1975–94 growth
decelerations were twice as frequent as accelerations:
29 percent versus 14 percent of all country-year observa-
tions in the Arbache-Page dataset.
The role of total factor productivity
Empirical work decomposing total factor productivity
(TFP) in Mauritius during different time periods shows the
strong role of factor accumulation. As in many countries in
East Asia, the principal drivers of growth in Mauritius have
been capital and labor accumulation, with TFP growth
making a significant but varying contribution.
In a paper using a growth accounting framework analy-
sis for Mauritius and highlighting the performance of
the 1980s and 1990s, Subramanian and Roy (2001) uncover
diverse performance in the two periods. For the period
1982–99, productivity growth in the EPZ sector was
3.5 percent, substantially more than double the 1.4 percent
for the economy as a whole. In the first period (1982–90),
economic growth was rapid and driven predominantly by
the growth of inputs—capital and labor—which together
accounted for more than 80 percent of the annual average
rate of GDP growth of 6.2 percent. During this period,
employment growth averaged 5.2 percent a year, reflecting
a sharp decline in the unemployment rate from 20 percent
of the total labor force in the early 1980s to 3 percent in the
late 1980s. TFP growth was respectable at more than 1 per-
cent, but capital accumulation was the main driver of
growth. The authors find that economic growth from 1991
to 1999, however, was driven less by capital accumulation
Southern Africa (COMESA) countries over the same years.
The TFP change in Mauritius resulted from economic
reforms and human capital improvements.
Structural transformation
Over time the sectoral composition of the Mauritian econ-
omy has changed profoundly. Between 1976 and 2010 the
share of primary-sector production declined from 23 per-
cent of the overall economy to 6 percent, while the second-
ary sector (including manufacturing, electricity, water, and
some construction) increased from about 23 to 28 percent
(with manufacturing making an even bigger jump). The ter-
tiary sector, which includes tourism and financial services,
grew from just over 50 percent to nearly 70 percent of GDP
(figure 5.3). Projections by the Mauritian government sug-
gest further expansion of the tertiary sector as a share of the
economy in the future. In general, the share of manufactur-
ing output increased in the 1980s because of the presence of
EPZs but stagnated as the sector faced adjustment in the
1990s and 2000s. Table 5.3 shows patterns for specific indus-
try groups and subgroups between 1990 and 2010. Sugar,
which represented more than 20 percent of Mauritius’s
GDP in 1976, accounted for approximately 4 percent of
GDP in 2009. At a disaggregated sectoral level (whether at
current prices or constant prices), there has been a strong
structural change, with the decline in sugar, a rise of finan-
cial services and real estate, and a mixed performance of tex-
tiles as the service sector has strengthened (figure 5.4) By
the 2000s the service sector had become the dominant fea-
ture of the economy in terms of contribution to output.
This noteworthy structural transformation has helped the
country deal with decreasing returns to scale of capital
accumulation at the sectoral level. The various economic
pillars in Mauritius have thus contributed to growth and
mitigated output volatility.

94 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY

 Table 5.2 Total Factor Productivity Decomposition and Contribution to Growth
Contribution to the Growth Rate
Growth rate Capital Labor TFP
Economy/ Years (percent) (percentage points) (percentage points) (percentage points)
Mauritius 1980–1990 6.3 3.0 2.0 1.3
1991–2000 5.6 2.5 2.1 1.0
2001–2007 4.1 2.5 0.9 0.7
1980–2007 5.3 2.6 1.7 1.0
COMESA 1980–1990 3.1 1.5 1.7 –0.2
1991–2000 2.4 0.7 1.6 0.1
2001–2007 4.5 2.1 1.5 0.9
1980–2007 3.3 1.5 1.6 0.3

Source: Rojid and Seetanah 2009.

Note: Percentages are rounded to nearest tenth.

Figure 5.3 Sectoral Composition of GDP in Mauritius, Poverty levels in Mauritius have also fallen significantly.
1976–2010 In 1975, 40 percent of Mauritian households were below the
presumed poverty line, according to the government’s
70 Central Statistics Office.8 By 1991/92 the proportion had
60 fallen to 11 percent, and by 2010 absolute poverty was less
50 than 2 percent. The poverty decline was achieved with no
Percent of GDP

40
exacerbation of inequality. Significant improvements in
gender equality have resulted from the massive inflow of
30
women into the labor market starting in the 1980s. The
20 country recently ranked 11th of 102 countries in the OECD
10 Social Institutions and Gender Index. Mauritius’s housing
0 stock has also improved, both in quality and quantity, as a
1976 1980 1985 1990 1995 2000 2005 2010 result of government investment. Finally, Mauritius has
Primary Secondary Tertiary developed a sophisticated pension system covering retire-
ment benefits and general social security.
Source: Mauritius authorities; World Bank 2009.

Improvement in human development indicators
In contrast to many Sub-Saharan African and comparator
economies, rapid economic growth in Mauritius has
occurred in parallel with substantial improvement in
human development indicators and a decrease in income
inequality. Life expectancy at birth, for example, increased
from 62 years in 1970 to 73 years in 2008, while infant
mortality dropped from 64 per 1,000 live births in 1970 to
15 in 2008 (figure 5.5). The Gini coefficient, a measure of
income inequality in which 0 represents perfect equality
among households and 1 represents perfect inequality,
declined from 0.5 in 1962 to 0.37 in 1986–87 and was stable
at 0.38 in 2008. Following heavy government investment in
education improvements in the 1980s and 1990s, primary
school enrollment rates reached very high levels, averaging
more than 90 percent in the 1990s and 2000s, although chal-
lenges remain.7
MACROECONOMIC MANAGEMENT
Prudent, proactive fiscal policy
A hallmark of economic management in Mauritius in recent
decades has been prudent fiscal policy, which has helped
maintain macroeconomic stability and contributed to
growth. Fiscal policy has focused on ensuring that spending
remains linked to resource availability. While there have been
fiscal imbalances, there is no history of the government bor-
rowing from the central bank or from aid agencies. The
strong growth in the 1980s led to a decrease in recourse to
foreign financing. The budget deficit, which was at 3 per-
cent in 1983, turned into a budget surplus by 1987, as
current expenditures shrank from 26 percent of GDP to 21
percent in 1987 while government revenues remained flat,
around 24 percent over the same period. On the expenditure
side, the government withdrew from subsidies on food items
and kept the wage bill under control during that period.

CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 95

 Table 5.3 GDP by Industry Group at Current Basic Prices, 1990–2010
(percent)
Industry group 1990 1995 2000 2005 2010a
Agriculture, hunting, forestry, and fishing 12.9 10.4 7.0 6.0 4.3
Sugarcane 8.0 5.7 3.6 3.2 1.6
Other 4.8 4.6 3.4 2.8 2.7
Mining and quarrying 0.2 0.2 0.2 0.1 0.0
Manufacturing 24.4 23.0 23.5 19.8 19.1
Sugar 3.4 1.6 0.8 1.0 0.5
Food, excluding sugar 0.0 — 4.1 5.1 6.6
Textiles 0.0 — 12.0 6.7 5.2
Other 0.0 — 6.6 7.0 6.8
Electricity, gas, and water supply 1.5 2.4 1.7 2.1 2.5
Construction 6.7 6.4 5.6 5.6 7.1
Wholesale and retail trade; repair of motor vehicles,
motorcycles, and personal and household goods 13.0 12.8 12.2 12.1 11.9
Wholesale and retail trade 12.6 12.3 11.7 11.4 11.1
Other 0.4 0.5 0.5 0.7 0.8
Hotels and restaurants 3.9 5.1 6.5 7.7 7.5
Transport, storage, and communications 10.4 11.4 13.0 12.6 10.8
Financial intermediation 4.9 6.5 9.7 10.3 12.3
Insurance 1.5 2.1 2.3 2.9 2.9
Banksb 0.0 4.4 6.6 6.2 8.0
Other 0.0 — 0.8 1.2 1.4
Real estate, renting, and business activities 8.9 8.5 8.9 10.2 11.9
Owner-occupied dwellings 6.4 5.3 4.5 5.0 5.0
Other 2.5 3.2 4.4 5.2 6.9
Public administration and defense; compulsory social security 6.4 6.7 6.7 7.1 6.5
Education 4.1 4.4 4.5 4.8 4.6
Health and social work 2.5 2.8 3.0 3.4 3.8
Other community, social, and personal service activities
and private households with employed persons 2.0 2.8 3.3 3.7 4.7
Financial intermediation services indirectly measured (FISIM) –1.8 –3.3 –5.7 –5.5 –7.0
GDP at basic prices 100.0 100.0 100.0 100.0 100.0
Manufacturing industries previously
operating with an EPZ certificate 11.9 11.4 11.9 7.4 6.7

Source: Central Statistics Office, Government of Mauritius.

Note: Figures may not add up to the totals due to rounding. — = not available. Data are at current prices.
a. Forecast figures.
b. For years 1991 to 1996, figures for other financial intermediation are included in banks.

Figure 5.4 Major Economic Sectors in Mauritius,

Similarly, in the 2000s, the government built up reserves
1976–2009
that allowed it the freedom to expand fiscal policy in the
100
aftermath of the 2008–09 global financial crisis. Moreover,
the stock of international and domestic debt has remained
80 well below an unsustainable threshold (total public debt
Percent of GDP

60 was projected to be 60.4 percent of GDP at the end of

2010). Excluding a period in the early 1980s, current
40
expenditures have never much exceeded 20 percent of GDP
20 and have been used mostly for wage policy, with a small
amount devoted to food subsidies. Taken together, the
0
1976 1980 1985 1990 1995 2000 2005 2009 composition of current expenditure has been mostly
Sugar Textile Construction
oriented in recent decades to the outlays for wages and sub-
sidies, but there has been little expansion of the federal
Hotels and restaurants Financial intermediation Real estate
apparatus as a share of GDP because fiscal profligacy has
Source: Mauritius authorities; World Bank 2009. been an anathema to the Mauritian policy makers. In the

96 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY

first four years of the 1980s, the ratio of current expendi- trade taxes. The diversified stream of revenue helped the
ture to GDP was close to 25 percent but it steadily dropped fiscal system absorb shocks and provide stable revenue
to around 20 percent by the mid 1980s, and that ratio has flows to the government.
stayed relatively constant, in contrast to that of many devel-
oping countries. Capital expenditures, which have averaged Monetary policy as an anchor
less than 5 percent of GDP since independence, have been for economic growth
used productively to invest in infrastructure—especially
In tandem with fiscal policy, monetary policy in Mauritius
roads—and to provide a necessary operating environment
has helped anchor economic growth and ensure competi-
for EPZs.9 Mauritian policy makers have been remarkably
tiveness. Since its creation in 1967, the Bank of Mauritius
fiscally prudent to avoid any macroeconomic instability. In
has been concerned with ensuring the competitiveness of
sum, fiscal policy has helped lay the foundation for manage-
the country’s export sectors and, secondarily, with price
ment of volatility and robust growth.
stability. In important ways, a series of exchange rate deci-
In terms of both revenue management and expendi-
sions early on had ripple effects on economic activity. As
tures, fiscal policy in Mauritius has been proactive. Inter-
part of the liberalization program, the Mauritian rupee
national trade taxes anchored the system’s revenue
(MUR) was devalued by 30 percent in 1979 and read-
system, accounting for close to 50 percent of GDP during
justed by 20 percent in 1981, when the rupee was officially
the 1970s, 1980s, and 1990s. High import tariffs and
delinked from the IMF’s special drawing right and pegged
export levies on sugar helped give the government
to a trade-weighted basket of the currencies of its major
resources during the early years, although the introduction
trading partners. By the mid-1980s the Bank of Mauritius
of a value added tax (VAT, which replaced the sales tax) in
was intervening to smooth currency fluctuations, and the
1998 has played an important role in improving tax buoy-
country has maintained a managed float since the mid-
ancy at the level of direct and indirect taxes and has
1990s. Figure 5.6 shows the path of the real effective
allowed the fiscal system to evolve.10 As a result, the tax
exchange rate (REER) of the rupee (base year is 1970 = 100).
system has not been affected by import tariff liberaliza-
The results of the managed float exchange rate regime have
tion in recent years. In addition, in 2007 the implementa-
been favorable to the economy, with the trade-weighted
tion of a flat tax of 15 percent on corporate and personal
REER depreciating on average by more than 5 percent a
incomes streamlined tax administration. By 2008 tax rev-
year from 1981 to 2007.11 Although inflation has been
enue amounted to 19 percent of GDP, of which 20 percent
higher in other Sub-Saharan African countries than in
came from income taxes, more than 35 percent from the
Mauritius, the central bank allows the accommodation of
VAT, and slightly less than 5 percent from international
these inflation differentials by letting the nominal exchange
rate slide in order to keep the REER competitive. The pol-
icy has worked well in terms of being flexible and being
Figure 5.5 Social Indicators in Mauritius, 1970 and 2008 sensitive to the country’s export sector.

80

70
Figure 5.6 MUR Real Effective Exchange Rate,
60 1981–2009
50
Number

Index (base year 1970 = 100)

130
40
120
30 110
20 100
10 90

0 80
Life expectancy (years) Infant mortality (per 1,000) 70
81
83
85
87
89
91
93
95
97
99
01
03
05
07
09

1970 2008
19
19
19
19
19
19
19
19
19
19
20
20
20
20
20

Source: World Bank 2010, UNICEF. Source: IMF.

CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 97

 As inflation fell in other parts of the developing world
in the 1990s, Mauritius adopted an informal inflation tar-
geting approach. Over time, monetary policy in Mauritius
has evolved from a strong reliance on direct monetary
instruments, such as credit ceilings, to a gradual intro-
duction of market-based instruments such as weekly
auctions of Treasury and Bank of Mauritius bills. In
practice, monetary management has been oriented to
ensure a positive interest rate differential relative to major
currencies while reacting to domestic inflation when it is
above a certain level. Interest rate policy has been used
successfully to provide savers with positive real rates of
return in order to mobilize domestic capital and to help
sterilize excess liquidity. As a result, national savings has
consistently exceeded investment expenditure since the
mid-1980s.12
Effective response to economic shocks
A tribute to Mauritius’s successful macroeconomic manage-
ment can be seen in the response to the global financial
crisis and the various shocks that preceded it. The combi-
nation of the phasing out of the Multifibre Arrangement
starting in 2004, the reduction in sugar price guarantees
from the European Union starting in 2006, and dramatic
increases in world commodity prices (especially for food
and fuel) had already started to act as a brake on the coun-
try’s growth trajectory and current account and fiscal
positions when the financial crisis began. The financial cri-
sis hit the small, open economy of Mauritius hard. The
country experienced a sharp decline in demand for tourism
and textile exports (figure 5.7). Labor-intensive sectors,
especially construction and textiles, contracted as the crisis
made its way to Mauritius.
As a result of the crisis, growth slowed to 4.2 percent in
2008 and to less than 2 percent in 2009, while unemploy-
ment increased from 7.2 percent in 2008 to 7.7 percent in
2009. Through prudent macroeconomic management,
however, international reserves continued to expand and
the fiscal deficit was kept below 5 percent of GDP. Moris-
set, Bastos, and Rojid (2010) find that the country’s
resilience to the shocks derived from a combination of four
factors: reforms to sustain long-term growth, which
accelerated in 2006; a timely, targeted, and temporary short-
term response to the crisis; institutional arrangements to
face the crisis that promoted private sector collaboration;
and strong relationships with development partners.
In response to the global financial crisis, Mauritius
passed a fiscal stimulus and monetary easing package of
98 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY
Figure 5.7 Trade and Tourism Growth, 2006–09
40
30
20
Percent
10
0
–10
–20
–30
2006 2007 2008 2009
Exports of goods Tourism receipts Imports of goods
Source: IMF 2010.
about 5 percent of GDP over 2009–10. The plan focused on
infrastructure spending, providing financial relief to the
firms hit hardest by the global crisis, and social and job
protection measures. At the same time, the government
introduced offsetting measures that are expected to bring
the primary budget into surplus by the end of 2011, and a
public debt management act was passed limiting public sec-
tor debt to 60 percent of GDP, with a goal of reducing it to
50 percent by 2013. The government has also been using
special “rainy day” funds to help finance the stimulus. These
funds had been prudently put aside in previous financial
years (to the tune of 3 percent of GDP), reducing current
debt financing needs. Monetary policy in Mauritius was also
loosened, the key discount rate at the central bank was cut
by 250 basis points, and reserve requirements were reduced.
In addition to easing fiscal and monetary policy,
Mauritius introduced measures to assist the private sector
in the wake of the crisis. Firms facing liquidity difficulties
were given temporary financial relief (conditional on credi-
ble restructuring plans), with costs shared by banks, the
government, and the firm’s shareholders.13 A tax suspension
program was also introduced for the labor-intensive and
vulnerable tourism, construction, and real estate sectors.
A TALE OF THREE SECTORS: SUGAR, EPZS,
AND TOURISM
At various times in Mauritius’s history, sugar, EPZs, and
tourism each have been a mainstay of the economy, and the
interaction among the three sectors has been essential for
the country’s economic take-off. As a small, open economy
with a high ratio of trade of goods and services to GDP
(averaging more than 50 percent from 1970 to 2010),
Mauritius has long been well positioned to embrace
globalization. Over the years Mauritius has used trade
policy as a means both to protect domestic industry and
to launch export growth.
The sugar sector
In its heyday in the 1970s the sugar sector in Mauritius
accounted for close to one-third of employment, one-third
of export earnings, and one-quarter of GDP. Through smart
negotiations and building on a preexisting relationship with
the United Kingdom, Mauritius succeeded in obtaining
preferential treatment from the European Economic Com-
munity (EEC) through the Sugar Protocol of the Lomé
Convention in 1975, under which it received more or less
free access for its sugar exports to the EEC.14 Mauritius sold
its sugar to the EEC at a premium—three times the interna-
tional market price, on average. For years Mauritius’s export
quota was fixed at more than 500,000 metric tons annually,
the largest quota share among African, Caribbean, and
Pacific (ACP) countries. Even with these international trade
agreements in place, however, the collapse in international
sugar prices in the mid-1970s hit Mauritius’s sugar sector
hard, leading to balance of payments difficulties and
recourse to external assistance. Nevertheless, by 2005, in the
aftermath of the World Trade Organization’s ruling that the
above-market prices paid to sugar producers constituted
unfair trade, the European Union ended the preferential
deals by slashing sugar prices.
Import substitution industrialization (ISI)
and restrictive trade policies following
the colonial period
Following the establishment of a development certificate
scheme by colonial authorities in 1964 to promote import
substitution industrialization and provide incentives for
local manufacturers through concessions and tariffs, Mauri-
tius invoked a series of protectionist measures in an effort to
develop local industry. Besides providing government with
some revenue and helping to incubate entrepreneurial tal-
ent, these policies, which were concentrated in the EPZs,
had a marginal impact on the local economy. Subramanian
and Roy (2001) find that Mauritius’s trade policy was highly
protective during the 1970s and 1980s. In 1980 average
effective protection exceeded 100 percent, although this fell
to 65 percent by the end of the 1980s. During the 1970s and
1980s, there were extensive quantitative restrictions in the
form of import licensing, which covered nearly 60 percent
of imports, and an extensive system of exemptions and con-
cessions. Protection was especially high in the clothing,
footwear, furniture, and rubber sectors, all of which had
tariffs above 50 percent, while tariffs for electronics and
plastics averaged more than 40 percent. Corporate taxes
were also very high, and bureaucracy was quite heavy. Pro-
tectionism faded through the course of the 1980s, however,
and by the early 1990s import licensing was eliminated on
all but a limited list of items subject to health, sanitary, or
strategic controls, while export licensing was abolished for
most products.
The rise of EPZs in the 1970s and 1980s
Having studied the success of export processing zones in
East Asia, a group of visionary policy makers in Mauritius
put forth the idea that the country’s small economic size
and distance from large developed markets presented a
potential opportunity to develop an export-oriented textile
industry. In 1970 Mauritius passed the Export Processing
Zone Act, which provided powerful incentives to manufac-
turers that catered to foreign markets. The EPZ was not
restricted to one physical location but was envisaged as a fis-
cal regime that encompassed the entire island, and the idea
was to develop a fast-track approval for all administrative
procedures in the EPZ. Ambitious and well-crafted legisla-
tion provided the underpinnings for the new regime. Key
components of the new legislation included protective
import duties and quotas for infant industries, suspension
of import duties on materials and equipment for industrial
use that were not locally available, rebates of import duties
on other raw materials and components for specified indus-
tries, duty drawback schemes, and favorable long-term
loans. The granting of duty-free inputs for manufactured
exports was key in expanding Mauritius’s export competi-
tiveness on world markets, while tax incentives helped sub-
sidize exports. Firms within EPZs also benefited from the
availability of relatively cheap labor, drawn from unem-
ployed workers and women who were outside the labor
force at the time. According to interviews with textile exec-
utives located in the EPZs, 80 percent of workers in the EPZs
in the 1980s were women. The rate has decreased somewhat
in the 1990s and 2000s, but more than 60 percent of the
workers in the zones are women. The lower wages that were
paid to the workers in the EPZs in the early years allowed
the firms to accumulate capital and reinvest the earnings into
the firms’ expansion.15 However, over time the wages in the
EPZ became higher than those in the non-EPZ economy.
CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 99
 Also important, Mauritius did not restrict EPZs to one geo-
graphical location, and the government invested heavily in
the infrastructure needed to set up EPZs. Finally, it is
important to note that the government provided strong
institutional support for marketing EPZ products.
By the 1980s EPZs had exceeded the expectations of
even visionary policy makers in Mauritius. The zones
accounted for more than 60 percent of Mauritius’s gross
export earnings and employed one-third of the Mauritian
labor force. Goods produced in EPZs more than tripled as
a share of GDP between 1980 and 1988, from 4 percent to
more than 14 percent (figure 5.8). More people worked in
EPZs than in the agricultural sector by the end of the 1980s.
The growth rate of the EPZs’ value added was close to 30
percent annually between 1983 and 1988. Most of the
goods produced in EPZs were exported to Europe under a
preferential regime. Notably, there was significant interac-
tion between the sugar sector and EPZs. Much of the
start-up capital for EPZ firms, as well as technical and
managerial expertise, came from the well-established sugar
companies in the aftermath of the sugar boom in the 1970s.
Together with the sugar sector, the textile sector provided
the capital accumulation that allowed Mauritius to
decrease reliance on foreign capital and start down the path
to becoming a middle-income economy.
Several additional factors contributed to Mauritius’s
success with EPZs. First, the country took advantage of the
depth of EU and U.S. demand for textiles and apparel,
which provided a solid base for expansion. Second,
Mauritius’s timing was good, as its initial entry into the
U.S. market got a boost from investors based in Hong Kong
SAR, China, who were seeking to move capital and factories
out of Hong Kong SAR, China, in anticipation of the 1997
Figure 5.8 Value of Mauritian Exports and Imports,
1980–2009
80
70
60
50
40
30
20
10
0 on raw sugar for all countries has thus been a shock for the
1980 1990 2000
Exports/GDP Exports of goods/GDP
Exports of services/GDP Imports of goods and services/GDP
Source: Mauritian authorities.
100 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY
reunification with China. Third, there was a strong political
stability, proper governance, and a clear legal and institu-
tional demarcation between the ISI and the EPZ regimes.
Fourth, quotas on Asian textile exports into Western
markets led investors to look to alternate production coun-
tries. Those investors brought capital, marketing networks,
and technological know-how to Mauritius’s nascent textile
sector. Greenaway and Milner 1989, however, find that the
decision to grant Mauritius trade preferences in garments
through the Multifibre Arrangement was more important in
giving the country privileged access to developed markets
relative to established Asian producers. Fifth, EPZs bene-
fited from the inflow of local capital and of indigenous
managerial capacity, which had partly been incubated
under import substitution policies. Bheenick and Schapiro
(1989) find that local participation in EPZ equity in Mauri-
tius was roughly half, a much higher ratio than in EPZs in
other developing countries. And finally, Mauritius benefited
from a strong entrepreneurial class and a trainable labor
force, that together with the political class, had a strong
ownership in the success of the EPZs. In turn, the job
opportunities offered by EPZs played a significant part in
unemployment falling from 20 percent to less than 5 per-
cent between the mid-1970s and 1990. Over time the EPZs
have helped the country’s exports match the imports,
although the proportion of service exports has increased in
recent years as the economy has changed (see figure 5.8).
Post-EPZ economic drivers: Tourism, business
process outsourcing, and financial services
Although EPZs brought dramatic economic improvements
to Mauritius, the textile and apparel sectors have met chal-
lenges in more recent years. Not least of these was the phas-
ing out of the Multifibre Arrangement, which started in 2004
and led to a contraction of 30 percent in value added of the
products produced in EPZs. The number of Mauritians
employed in EPZs fell by about 25,000 people between 2005
and 2010.16 In parallel, the European Union’s reduction of
sugar prices by more than 50 percent starting in 2005 and
continuing onto 2010 was a significant blow to the Maurit-
ian economy, given that from 1975 to 2005, about 90 percent
of its sugar production was exported to the European mar-
ket. The European Union’s decision to end price guarantees
2009 $10 billion Mauritian economy. In short, both the sugar and
textile and apparel sectors are in the process of adjustment in
order to remain globally competitive, with the textile sector
making inroads into fully integrated activities (for example,
spinning and weaving) and higher-end manufacturing and
the sugar sector increasing its focus on refined sugars.
As the sugar and EPZ sectors in Mauritius have struggled
in recent years, the tourism sector has expanded rapidly,
backed by a master plan reflecting the government’s dislike
for mass tourism and high-rise buildings. Public sector efforts
to expand tourism have been complemented by the promo-
tional activities of the hotels and by Air Mauritius. Indeed,
Mauritius’s combination of beautiful beaches, a multieth-
nic society, and excellent hotels dotting the coastline has
been quite effective in attracting tourists. According to the
Mauritius Chamber of Commerce, tourist arrivals reached
240,000 in 1988, 400,000 in 2000, and 900,000 in 2008. An
estimated 1 million tourists were expected to visit the
country in 2010. Tourism is also among the strongest for-
eign exchange–earning sectors of the Mauritian economy.
In addition to tourism, the government of Mauritius has
also encouraged diversification of the economy into business
process outsourcing (BPO), financial services, and informa-
tion technology. According to government figures, the BPO
industry has been growing 70 percent a year and is currently
worth $1.6 billion, employing more than 100,000 people.
Offshore banking was introduced in 1988 as a first step
toward developing Mauritius into an international financial
center, and the offshore sector is now emerging as a growth
vehicle for the economy.17 Development of the information
technology sector, meanwhile, is intended to transform
Mauritius into “a cyber island” by creating a high-tech,
multi-story tower that provides a home to companies from
all over the world to set up BPO operations, including data
management, e-commerce, and call centers. The government
has also encouraged use of Mauritius as a transshipment
center and a re-export base and, more recently, as an inter-
national medical hub and regional knowledge center. As a
result of these efforts, the Ministry of Finance finds that the
services sector showed consistent growth over 2006–09, with
financial intermediation growing by 10.1 percent in 2008.
Lessons of the three sectors
The story of the three sectors in Mauritius offers a number
of lessons. Most important, the constant search for new
drivers of economic growth reflects a desire by policy mak-
ers to adapt to the future rather than wait for and respond
to shocks. The central lesson here is that the public sector
can effectively formulate and implement sectoral policies to
stimulate the private sector. In the case of sugar, the
protocols were signed and the private sector rushed into the
activity. In the case of EPZs and tourism, a well-articulated
policy framework led to a strong private sector response.
And in all cases, the government acted as a facilitator of pri-
vate sector expansion.
DYNAMIC INSTITUTIONS AND ADAPTABILITY
TO CHANGE
Aside from its success in macroeconomic management, one
of the keys to Mauritius’s economic success has been its rich
web of trade links, effective institutions, and history of
public collaboration. In surveys of institutional quality,
Mauritius repeatedly ranks high relative to comparator
countries, particularly in measures of governance, rule of
law, and control of corruption. The combination of political
stability, democratic legacy, rule of law, and quality of judi-
cial institutions sets Mauritius apart from many Sub-
Saharan African countries. Moreover, a set of informal and
formal mechanisms guide the interaction between the pub-
lic and the private sector, with the result that the private sec-
tor plays a seminal role in the policy formulation process
(all Mauritian delegations to international organizations,
for example, have a private sector member). Indeed, coop-
eration between the public and private sectors has a long
history in Mauritius.
For decades, Mauritius served as a trade hub for Chinese
and Indian traders and an entrepôt for shipping across the
Indian Ocean, and its inhabitants were known for their
entrepreneurship. Over time those trade links coalesced into
formal trade associations and entities, some of which came
to be represented at the political level. At the same time,
Mauritius has been quite effective at promoting its trade
links on the international level.
The importance of forging consensus
The search for consensus is one of the remarkable features
of the Mauritian political economy. Essentially, the Maurit-
ian state is modeled on the British system of government,
with a cabinet headed by a prime minister and a legislative
assembly serving as the law-making body. As in the United
Kingdom, the system of government in Mauritius is based
on the principle of separation of power between the legisla-
ture, the executive, and the judiciary. Gulhati and Nallari
(1990) argue that since no single political party has ever
secured a majority in the assembly, which would allow it to
form a government on its own, there has always been a need
to work together across party lines, putting a distinctive
stamp on economic policy process. Despite being based on
loose agglomerations of ethnic and economic interests,
CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 101
 political parties in Mauritius have not been vehicles for eth-
nic separation. To the contrary, political parties have long
recognized that building consensus is necessary to avoid
adverse economic effects in a small economy.
The nexus between the public and private sectors
A paramount role in state-business relations in Mauritius is
played by the Joint Economic Council (JEC), which occu-
pies a central place in the country’s institutional landscape
and represents an umbrella association of a number of
sector-specific groupings. As such, it carries a certain
amount of institutional weight, meeting with the prime
minister on a regular basis and providing input on major
policy decisions. Being funded entirely by the private sector,
it also has a degree of financial autonomy. The overarching
goal of the JEC is to ensure private sector representation in
all key government economic decisions. It also ensures that
its members’ ideas are conveyed to political leaders.
Another example of public-private sector cooperation is
the establishment of the EPZs, which would not have been
possible without support from several key public sector
institutions. Central among these were the Mauritius Export
Development and Investment Authority (MEDIA), the
Export Processing Zone Authority (EPZDA), and the Devel-
opment Bank of Mauritius (DBM). MEDIA was formed in
1985 as a public trade and investment promotion agency
(with some private sector membership) and was a pivotal
institution behind the country’s drive for export growth and
industrial upgrading. Providing overseas marketing support
for exports and arranging buyer-seller meetings, it helped
explore niches for Mauritian garments in European and
American markets. Formed in the early 1970s, the EPZDA
helped represent the interests of firms in the EPZs, while the
Development Bank of Mauritius provided much of the
credit and start-up capital for the economy as it was taking
off from its narrow monocrop base. More recently, the Board
of Investment has played a role in helping to promote Mau-
ritius as an international investment, business, and service
center, providing counseling on investment opportunities in
Mauritius and helping in setting up businesses.
BUSINESS CLIMATE AND INVESTMENT
Alongside successful trade policy and adaptability, another
major reason for Mauritius’s economic success has been its
business climate and incentives for foreign companies to
locate there. Mauritius has no capital controls, a relatively
stable currency, a low flat corporate tax rate of 15 percent,
102 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY
and a large number of double taxation avoidance agree-
ments; together, these attributes sometimes make Mauri-
tius more attractive than larger financial sectors for busi-
nesses. International rankings consistently give Mauritius
high marks for business and investment climate. The main
lesson from Mauritius in this regard is that geography is
not destiny and that policies to improve investment cli-
mates can have large positive multiplier effects.
Business climate improvement in recent years
The World Bank’s Doing Business 2010 ranks Mauritius as
the best country in which to do business in Africa. Overall,
Mauritius is ranked 20th of 183 countries included in the
2010 survey, up from 24th of 183 in 2009. Currently,
Mauritius is among the top-performing developing coun-
tries in starting a business, paying taxes, and protecting
investors, and it has been a consistent reformer since it
began being included in Doing Business in 2005. More
broadly, Mauritius has taken steps to improve business
facilitation since the late 1990s. In earlier years, while the
private sector was thriving in certain sectors such as sugar
and textiles, the climate for domestic firms in the non-EPZ
sector was unfavorable. A study by Lall and Wignaraja
(1998) found several major obstacles for enterprises oper-
ating in Mauritius at the time: high interest rates; heavy
bureaucratic procedures resulting in delays in obtaining
foreign investment approvals; difficulty getting loan
approvals from the Development Bank of Mauritius; delays
in receiving refunds on import duties; difficulty obtaining
work permits for foreign technical staff; lack of access to
finance for small enterprises; and high sea-freight costs.
Government reforms have helped alleviate all of these
obstacles in the years since.
One major piece of legislation, the Business Facilitation
Act 2006, provided a new, streamlined legal framework for
business operations in Mauritius. The legislation facilitates
doing business and acquisition of properties by foreigners
and, among other things, enables small enterprises to start
business activities within three working days. As a result, the
private sector and foreign investors can more easily venture
into new sectors, such as real estate and financial services, in
which growth rates are eclipsing traditional sectors.
Between 2006 and 2010, the share of private sector invest-
ment, mainly in infrastructure projects such as commercial
and office buildings, hotels, and resorts, grew to account for
more than 80 percent of gross domestic capital formation.
While infrastructure deficiencies and difficulties obtaining
credit remain for small firms, along with a lack of skilled
labor for larger firms, the thrust of the reforms has been to
alleviate these constraints.
The role of foreign direct investment
FDI inflows to Mauritius have increased rapidly in the past
several years, attracted by reforms such as the removal of
the tax on capital account transactions and the waiving of
the requirement that foreign investors need approval of the
Bank of Mauritius to carry out activity. In addition while
the corporate tax rate is a low 15 percent, foreign firms
receive a subsidy of close to 10 percent, leading to an effec-
tive tax rate of 5 percent. The country attracted more FDI
during 2004–07 than the cumulative stock of FDI during
the previous 25 years (figure 5.9). Importantly, FDI inflows
are accompanied by new business ideas, technologies, and
managerial skills. Most FDI inflows have gone to the hospi-
tality and tourism, property and real estate, banking and
finance, information technology, health, and education sec-
tors. The main FDI source countries are France, South
Africa, and the United Kingdom, although total FDI inflows
are equally divided between developed and emerging coun-
tries. Interestingly, Mauritius has also been a beneficiary
of a high inflow of FDI into India. Because of special tax
treatment given to investments that come through Mauritius
to India,18 Mauritius has become a quasi-tax haven for for-
eign funds invested in India, and currently, about 80–90 per-
cent of foreign direct investment into India flows through
Figure 5.9 FDI Inflows to Mauritius, 2002–08
12,000
10,000
Millions in local currency
8,000
6,000
4,000
2,000
0
02
03
04
20
20
20
Manufacturing
Hotels and restaurants
Real estate, renting
Source: World Bank 2009.
Mauritius through private equity, hedge, and mutual funds.
Under the current double taxation treaty between India
and Mauritius, capital gains on Indian shares that are held
by a Mauritian company are not subject to Indian tax laws
and rates, an issue that has been vexing to Indian regulators
and policy makers. Simply put, Mauritian companies are
taxed according to Mauritius tax laws, which are extremely
favorable, compared to Indian laws. There is some chance
that India will try to amend the treaty in 2011 to try to
capture more of the revenue from this activity. Neverthe-
less, FDI from Mauritius to India, which has been mostly
in the electrical equipment, cement, telecommunications,
and financial sectors, has helped Mauritius establish the
attributes needed to compete globally in high-value ser-
vice sectors.
THE ECONOMIC FUTURE OF MAURITIUS
The long-term challenge for the government of Mauritius is
to maintain its unique combination of resilience in the face
of changing economic circumstances and adaptability to
new paths for growth. A number of key lessons can be
learned from Mauritius’s experience. First, the forging of
consensus between the Franco-Mauritian business elite and
the Indian political elite has provided a sound foundation
for economic growth. Consensus building was mapped into
the management of ethnic interests while the country
opportunistically moved forward with growth strategies.
05
06
07
08
20
20
20
20
Wholesale and retail trade; repair of motor vehicles
Financial intermediation
Other
CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 103
 Even an occasionally contentious political environment has
not jeopardized the continuity of state policies and admin-
istrative stability. The combination of adaptable institutions
and a rich interface between the public sector and private
sectors has ensured effective economic policy. The impor-
tant role the private sector plays in the formulation of
economic policy, especially through the Joint Economic
Council, is relatively unparalleled in Africa.
Second, Mauritius benefited from pragmatic macroeco-
nomic management that was supportive of long-term growth
aspirations. The rents generated within the system during the
1970s and 1980s were used to finance capital accumulation
rather than consumption. The real exchange rate was kept
competitive, fiscal discipline was maintained, and debt bur-
dens were kept at respectable levels, while there was a willing-
ness to correct external and internal imbalances when needed.
Third, Mauritius recognized the benefits of economic
openness at an early stage, implementing effective sectoral
polices and building a good investment climate. The policy
framework was used to facilitate private and foreign invest-
ment, particularly in the textile and tourism industries. The
interplay between the relatively closed import substitution
industrialization on one hand and export-driven initiatives
on the other (particularly the EPZs) provides a fascinating
tale. In general, though, Mauritius’s openness, which
allowed it to successfully penetrate developed markets
through exports, has been unrivaled by countries in Africa
and the Middle East. Through smart tax policy, the country
has become a source of a large portion of FDI flows into
India. Throughout, Mauritius has demonstrated a capacity
to capitalize on good international relationships.
While its future success is not guaranteed, Mauritius has
proven that it has the right instruments to weather a range
of economic shocks. Its history is one of adaptability, inno-
vation, and anticipating global changes. Its combination of
good leadership, consensus building, sound macroeco-
nomic management, and positive policies for the private
sector will serve it well in the future.
NOTES
1. Careful empirical work by Subramanian (2009) shows
how initial conditions in Mauritius—especially the
income level, geography, and commodity dependence—
have hurt long-term growth, while favorable demography
and high levels of human capital have been mitigating
factors. For example, Mauritius is disadvantaged by being
at least 25–30 percent more distant from world markets
than the typical African country. Statistical analysis shows
104 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY
that on balance, the disadvantages outweigh the advan-
tages: initial conditions have slowed growth by about 1
percentage point a year relative to the average African
country and by nearly 2 percentage points relative to the
fast-growing developing economies of East Asia.
2. See Sachs and Warner 1995, 1997; Meade 1961; Rodrik
1999; Romer 1993; and Subramanian and Roy 2001, among
others.
3. Mauritius has three ethnic groups: Mauritian Creoles,
Indo-Mauritians, and Franco-Mauritians. The Creoles
(African) were brought to Mauritius as slaves to work for
owners of sugar cane fields. Indo-Mauritians came to
Mauritius as indentured laborers after slavery was abolished
in 1835 and eventually became the country’s political elite.
And Franco-Mauritians were the French who remained in
Mauritius after the British took over in 1810 to look after
their large sugar estates and other businesses, including
trading and banking.
4. Mauritius’s legislative system is based on a classic parlia-
mentary Westminster system. Legislative power is vested in
the National Assembly, which is composed of 62 elected and
up to 8 designated representatives. The four main current
political parties in Mauritius are the Labour Party (PTR),
the Movement Mauricien Militant (MMM), the Mauritius
Socialist Militant (MSM), and the Parti Mauricien Xavier
Duval (PMXD).
5. Although Mauritius’s public debt is relatively high, at
about 60 percent of GDP, most of it is held domestically by
the National Pension Fund and commercial banks.
6. Note that TFP calculations are very sensitive to the start
and end year.
7. One challenge Mauritius faces is the large proportion of
young people who are unable to access secondary education
because of the very competitive system for moving from
primary to secondary schools and who thus cannot con-
tribute to the skilled labor force required by the economy.
Almost 35 percent of students fail to pass the completion-
of-primary-education examination and drop out of the
school system at the age of 12.
8. Mauritius does not have a national poverty line; poverty
figures are derived by the Central Statistics Office using
census and survey data.
9. It should be noted that capital expenditures are under-
estimated in the Mauritian budget, because many public
investments take place through parastatals, while expendi-
ture classifications are detailed for the central government.
As a result, some capital expenditures may be classified as
current expenditures.
10. The VAT was introduced in Mauritius in September
1998, close to six months ahead of schedule, and its perfor-
mance has exceeded expectations.
11. The IMF (2008) finds that using the macroeconomic
balance approach and the single-equation equilibrium
exchange rate approach, the real exchange rate at the end of
2007 was broadly in line with its equilibrium value (as
determined by economic fundamentals).
12. Since the mid 1980s private sector investment has also
exceeded private sector savings.
13. Technically, the Mechanism for Transitional Support for
the Private Sector (MTSP) under the Additional Stimulus
Package (ASP) was a combination of equity support,
liquidity/working capital (including guarantees for bank
support), and asset purchases and swaps. Between mid-
December 2008, when the mechanism began functioning,
and September 2010, 11 companies had received assistance
under the MTSP, to which the government contributed
MUR 140 million (36 percent of the total MTSP support for
these companies; the remainder was provided by banks and
shareholder equity) in the form of debentures at 5 percent
interest.
14. The Sugar Protocol was negotiated as a condition of the
United Kingdom’s membership in the EEC to protect the
developing countries from which it had traditionally
imported sugar under the Commonwealth Sugar Agreement.
15. Firms within EPZs were subject to general labor laws
(including minimum wages) but were free to fire workers, to
demand compulsory overtime work, and to penalize work-
ers heavily for absenteeism.
16. Despite the challenges, the government has supported
the restructuring of EPZ textile firms in order to avoid a
possible collapse of the sector. These efforts include estab-
lishing the Textile Emergency Support Team initiative in
July 2003 and encouraging the National Productivity and
Competitiveness Council to carry out a diagnostic study of
textile firms to assess their cost structure and point to areas
in need of improvement.
17. At the end of October 2002, the number of companies
registered in the offshore sector reached 20,111. The Mauri-
tius Freeport, the duty-free zone in the port and airport,
aims at transforming Mauritius into a major regional distri-
bution, transshipment, and marketing center.
18. India has double taxation avoidance agreements with
approximately 65 countries, including France, Germany,
Japan, the United Kingdom, and the United States, although
Mauritius is the most preferred route for FDI inflows.
BIBLIOGRAPHY
Arbache, J., and J. Page. 2008. “Hunting for Leopards: Long-
Run Country Income Dynamics in Africa.” Policy
Research Working Paper 4715, World Bank, Washington,
DC.
Bheenick, R., and M. O. Schapiro. 1989. “Mauritius: A
Case Study of the Export Processing Zone.” EDI
Development Policy Case Studies 1, World Bank,
Washington, DC.
Greenaway, D., and C. Milner. 1989. “Nominal and Effective
Tariffs in a Small Industrializing Economy: The Case of
Mauritius.” Applied Economics 21 (8): 995–1009.
Gulhati, R., and R. Nallari. 1990. “Successful Stabilization
and Recovery in Mauritius.” EDI Development Policy
Case Studies. World Bank, Washington, DC.
IMF (International Monetary Fund). 2008. “Mauritius:
Selected Issues.” Washington, DC.
———. 2010. “Mauritius Article IV Consultation Staff
Report.” Washington, DC.
Lall, S., and G. Wignaraja. 1998. “Mauritius: Dynamizing
Export Competitiveness.” Commonwealth Secretariat,
London.
Meade, J. E., et al. 1961. The Economics and Social Structure
of Mauritius: Report to the Government of Mauritius.
London: Methuen.
Morisset, J., F. Bastos, and S. Rojid. 2010. “Facing Off a Man-
Made Disaster: Global Financial Crisis and Policy
Response in the Small Tropical Island of Mauritius.”
World Bank Note, Africa Region, Washington, DC.
Rodrik, D. 1999. The New Global Economy and Developing
Countries: Making Openness Work, London: Overseas
Development Council.
Rojid, S., and B. Seetanah. 2009. “Using Growth Accounting
to Explain Sources of Growth: The Case of COMESA.”
International Journal of Business Research.
Rojid, S., B. Seetanah, and R. Shalini. 2009. “Are State Busi-
ness Relations Important to Economic Growth: Evidence
from Mauritius.” DFID/University of Mauritius working
paper.
Romer, P. 1993. “Two Strategies for Economic Development:
Using Ideas and Producing Ideas.” In Proceedings of the
World Bank Annual Conference on Development Econom-
ics. Washington, DC: World Bank.
Sachs, J., and A. Warner. 1995. “Economic Reform and the
Process of Global Integration: Comments.” Brookings
Papers on Economic Activity, 1: 1–118.
———. 1997. “Sources of Slow Growth in African
Economies.” Journal of African Economies 6 (3): 335–76.
Subramanian, A. 2009. “The Mauritian Success Story and its
Lessons.” Research Paper 2009/36, United Nations
University–World Institute for Development Economics
Research, Helsinki.
Subramanian, A., and D. Roy. 2001. “Who Can Explain the
Mauritian Miracle: Meade, Romer, Sachs, or Rodrik?”
IMF working paper, WP/01/116, Washington, DC.
CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 105
 Wignaraja, G., and S. O’Neil. 1999. SME Exports and Public
Policies in Mauritius. Commonwealth Trade and Enter-
prise Paper. London: Commonwealth Secretariat.
World Bank. 1989. Mauritius Country Economic Memo-
randum, Washington, DC.
106 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY
———. 2009. Mauritius Investment Climate Assessment.
Washington, DC.
———. 2010. World Development Indicators Database.
Washington, DC.
 CHAPTER 6
Cotton Dependence in Burkina Faso:
Constraints and Opportunities for
Balanced Growth
Jonathan Kaminski
urkina Faso experienced more than a decade of
B
robust growth in gross domestic product (GDP)
following the devaluation of its currency in 1994.
Average growth of 6 percent a year—nearly double the
growth of the previous decade—was led largely by the cot-
ton sector, which accounted for about 60 percent of exports.
GDP per capita in Burkina Faso rose from $214 in 1997 to
$260 in 2007 ($430 in real terms). Reforms in the cotton
sector contributed to rapid growth of this sector, but
spillover in terms of structural transformation of the econ-
omy has been slow, especially because the increase in cotton
production was not productivity-based. While reforms that
led to management improvement and institutional upgrad-
ing in the cotton sector were necessary, they were not suffi-
cient to translate productivity increases in the cotton sector
into economic consolidation, structural transformation,
and diversification. Instead, the economy’s dependence on
cotton has exacerbated growth volatility and vulnerability to
exogenous shocks. Despite this, Burkina Faso’s economic
prospects remain positive due to record cotton prices in the
world market, a growing level of gold exports, effective gov-
ernment-led growth strategies, and continuing improve-
ment of the business environment.
ECONOMIC GROWTH IN BURKINA FASO
Burkina Faso’s average annual GDP growth rate accelerated
from an average 3.8 percent in the 1980s to 6 percent during
1994–2008. The agricultural sector, which accounts for
85 percent of Burkina Faso’s active labor force, has been the
main driving force of growth. Within agriculture, the cotton
sector has shown particular dynamism. The cotton boom
during the 2000s lifted cotton’s share of GDP from around
4 percent in the 1980s to 12 percent in 2008, contributing to
higher GDP growth (figure 6.1).
As cotton production in Burkina Faso posted unprece-
dented growth in the 2000s, the share of cotton earnings in
export revenues shot up from less than 40 percent in the 1990s
to 85 percent in 2007. At the same time, increased export
dependency on cotton has exacerbated vulnerability to exoge-
nous shocks over the past decade, which was characterized by
a pattern of falling world cotton prices and rising input prices;
a decline in local profitability and farm productivity; and
poorly performing cotton firms that lack the ability, informa-
tion, and resources to adjust to evolving international markets.
Although cotton represents a large proportion of Burkina
Faso’s exports, its contribution of export earnings to GDP is
small (10 percent) and trade openness is limited. Most of the
country’s growth is from domestic demand.1 Limited export
earnings highlight the unsustainable growth path that Burkina
Faso is currently on. Although the country’s terms of trade
have recovered since the early 2000s, structural deficiencies
have led to persistent trade deficits (Savadogo 2009).
EFFECT OF GROWTH ON DEVELOPMENT
Sustained growth has had only a limited effect on poverty
reduction in Burkina Faso (World Bank 2009a). Despite
107
 Figure 6.1 Annual Changes in Cotton Production and GDP in Burkina Faso, 1980–2008

12 100

10 80

60
8
40
6
Percent

Percent
20
4
0
2
–20
0
–40

–2 –60

–4 –80
81

83

85

87

89

91

93

95

97

99

01

03

05

07
19

19

19

19

19

19

19

19

19

19

20

20

20

20
Percent change in GDP (left axis) Percent change in cotton production (right axis)

Source: World Bank 2010; FAOSTAT, SOFITEX.

sustained annual per capita GDP growth of about 2–3 percent
between 1994 and 2003, the national poverty headcount
index remained about 45 percent at the end of that period.
Granted, estimation of poverty measures has been subject to
controversies (including methodological), but no significant
effect of growth on poverty reduction has been found at the
aggregate level (Lachaud 2005, 2006; Grimm and Gunther
2004). Grimm and Gunther (2004) show that only cotton-
producing households increased their expenditures signifi-
cantly between 1995 and 2003, a finding that attests to the
pro-poor growth effect of smallholder cotton production.
Several studies find that the impact of Burkina Faso’s GDP
growth was mostly neutralized by its population growth,
which limited scaled-up investments in infrastructure,
education, and health.
Several pieces of research show evidence of poverty
reduction in the mid- to late 2000s: the incidence of poverty
fell from 46 percent in 2003 to 38.5 percent in 2007; the inci-
dence of poverty in rural areas fell from 52 percent to
44 percent over the same years. There is no evidence, how-
ever, that cotton-producing regions had lower incidence of
poverty than other rural areas (Grimm and Gunther 2004).
The apparent contradiction between pro-poor cotton
growth on one hand and the absence of rapid poverty
reduction in cotton-producing areas on the other reflects
both the effect of political and economic turmoil in Côte
d’Ivoire (which reduced remittances to Burkina Faso) and
some amount of poverty convergence across the West
African region as a result of the migration of less productive
and poorer farmers in cotton regions to more marginal
lands (Gray and Kevane 2001). In addition, the increase in
cotton production reduced poverty largely among house-
holds that were able to increase their per capita agricultural
incomes through more productive assets and factors. There
was no significant poverty reduction effect as a result of the
decline in cereal prices or spillovers into the local economy
led by higher domestic demand (Kaminski, Headey, and
Bernard 2009).
As global food and fuel price increases took hold in 2007,
the purchasing power of households in Burkina Faso
declined and the incidence of poverty began to rise again.
Cotton households did not escape these negative impacts,
and in subsequent years, negative shocks such as flooding
have hurt cotton-producing households even more. At the
same time, the deceleration in economic growth had a dra-
matic effect on poverty, alleviation of which requires higher
rates of economic growth, mainly in the rural sector (cotton
and livestock). Measures of the depth and severity of
poverty in Burkina Faso have also increased since 2007.
Although poverty is much higher in rural areas than in
urban areas, unemployment in Burkina Faso is largely an
urban problem that disproportionately affects women, youth,
and highly skilled workers.2 The urban unemployment rate
fell to 8.6 percent in 2007, down from 13.8 percent in 2003

108 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
(16 percent for the labor force ages 15–25, down from
30 percent in 2003). According to the World Bank (2009a),
the quality of jobs has not improved, however, and a large
proportion of the workforce remains employed in low-
paying jobs and jobs that do not provide social benefits.
Absorbing young graduates in professional jobs remains a
major challenge for the government in the coming years. A
new law that promotes more flexibility in the labor market
went into effect in 2008, but creation of new jobs has been
hampered by the economic slowdown.
Other indicators of social welfare and development
have exhibited encouraging signs over the past decade.
Education-related infrastructure has been scaled up, with
the number of classrooms increasing by 9 percent annually
since 2006. Primary school enrollment rates have also
risen, reaching 78 percent in 2009. Although regional dis-
parities are significant, gender disparities in enrollment
rates have narrowed. These results have been achieved
thanks to long-term policy frameworks, substantial
increases in public investment in education in the 2000s,
and partnerships with the private sector (which has con-
tributed about 15 percent of investment in education
infrastructure).
In the health sector, the prevalence of HIV/AIDS
has fallen, and progress has been made in containing
epidemic and endemic diseases. Both morbidity and
mortality rates decreased between the mid-1990s and
mid-2000s.
REFORM OF THE COTTON SECTOR
Since the early 1990s Burkina Faso’s dependence on cotton
has grown as a result of the implementation of institutional
reforms, which have brought new land and producers to
cotton production (Kaminski and Thomas 2009). Table 6.1
details the chronology of the reforms.
The cotton sector provides income for 15–20 percent
of the active labor force of Burkina Faso, supporting 1.5
million–2 million people. It is composed mostly of small
farms and smallholders, with a small number of large
farms led by rural elite. In 2005 Burkina Faso became the
leading West African producer of cotton, ahead of Mali,
producing 500,000 to 800,000 tons of seed cotton between
2005 and 2010. In 2006 and 2007, Burkina Faso was the
leading cotton producer and exporter among all African
countries.

Table 6.1 Chronology of Cotton Reforms in Burkina Faso, 1992–2008

Year(s) Development
1992–93 Formal commitment made by Societé Burkinabé des Fibres et des Textiles (SOFITEX), the national cotton parastatal
company, to let producers’ representatives participate in reform debate. Contrat-Plan Etat SOFITEX, in which the state
committed not to interfere with management of SOFITEX, established a plan to streamline accumulated debts of
producers and the parastatal.
1994 Laws pertaining to establishment of farmer groups amended.
1996–99 Free membership introduced in formation of groups of local cotton farmers; the groupements villageois were replaced by
market-oriented organizations (groupements de producteurs de coton) at the subvillage level, with implementation of new
local governance rules.
1996–2001 National cotton union (Union Nationale des Producteurs du Coton du Burkina, or UNPCB) is established with support of
l’Agence Française de Développement, the Burkinabe government, and SOFITEX, based on membership of local groups
and their integration into regional unions.
1998 Accord Interprofessionnel signed by SOFITEX, the state, UNPCB, donors, and financial consortium (Caisse Nationale du
Crédit Agricole; Banque Internationale pour le Commerce, l’Industrie, et l’Agriculture; Banque Internationale du Burkina),
replacing the Contrat-plan and defining the reallocation of responsibilities.
1999 State partially withdraws from the sector through partial privatization of SOFITEX; half of government’s share in SOFITEX is
transferred to UNPCB.
2000–06 Economic activities—including provision of cereal input credit; management assistance of cotton groups; and participation in
quality grading, financial management, and price bargaining—progressively delegated from SOFITEX and the government to
UNPCB. The state downsizes support of research and extension services.
2002–06 New players—including private input providers, new regional private cotton monopsonies (SOCOMA, FASOCOTON), and
private transport companies—begin operating in the sector.
2004–06 Interprofessional association (AICB) is established with cooperation of cotton farmers, banks, private stakeholders, the
government, and research institutes. Association of cotton firms (APROCOB) established to interact with UNPCB.
2006–08 Price-setting mechanism changed to better reflect world price levels; new smoothing fund managed by an independent
organization becomes operational in 2008.

Source: Kaminski, Headey, and Bernard 2009.

CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 109
 A partial equilibrium analysis based on estimation of coun- Figure 6.2 Seed Cotton Production in Selected
terfactuals suggests that cotton reforms explain two-thirds Sub-Saharan African Countries, 1980–2010
of the threefold increase in production from 1996 to 2006, 700

Average annual production of cotton

which enabled farmers to cultivate more land, raise their

seed (thousands of metric tons)

incomes, and improve food security (Kaminski, Headey, 600

and Bernard 2009). Figure 6.2 compares average production 500

of seed cotton in Burkina Faso and neighboring West
400
African countries with that in the Southern African coun-
tries of Zambia and Zimbabwe, in which conventional pri- 300
vatization and reforms were pursued (table 6.2). As shown 200
in figure 6.2, the volume of cotton production growth in
Burkina Faso has been substantially higher than in com- 100
parator countries since 2000. 0
Importantly, increased cotton production in Burkina

0
98

98

99

99

00

00

00

01
–1

–1

–1

–1

–2

–2

–2

–2
Faso in recent years was not based on significant productiv-

80

85

90

95

00

05

07

09
19

19

19

19

20

20

20

ity increases but rather on accumulation of factors such as
land, labor, and inputs. Rising cotton yields also hide the
negative dynamic effect resulting from the entry of less able
farmers and land in cotton production. Hence, stagnant
(even slightly increasing) yields mean that individual per-
formance increased over the last decade. Overall competi-
tiveness of cotton farmers and of the cotton industry
increased much more rapidly than in neighboring countries
and has been hampered only by low world market prices
(which include the depreciating effect of U.S. cotton subsi-
dies) and the appreciation of the euro.
Nevertheless, cotton-led growth faces several challenges.
Kaminski, Headey, and Bernard (2009) point out the lack of
sustainability of cotton reforms as well as the limitations of
the factor-driven Burkinabe economy, especially with regard
to the underlying environmental constraints to extensive
agricultural growth and the need for the intensification of
sustainable agricultural. The cotton sector also faces envi-
ronmental challenges, including the threat of soil impover-
ishment and emerging conflicts regarding access to land
(Gray and Kevane 2001). These problems notwithstanding,
the reform has led to upgraded institutional arrangements
between producers and other stakeholders and helped farm-
ers organize in more professional-oriented associations to
carry out a growing number of responsibilities (World Bank
2004). All these factors improved the overall management of
the sector. The reform has resulted in higher profit-sharing
for farmers, better extension services, fairer quality grading,
and better access to agricultural inputs with improved credit
repayment mechanisms (Kaminski 2007).
The Burkinabe reform model is unique in Sub-Saharan
Africa in that it addresses government failures and local
realities within the current institutional framework,
20
Benin Burkina Faso Côte d'Ivoire
Mali Zambia Zimbabwe
Source: Kaminski, Headey, and Bernard 2009.
adopting reforms using a cautious, piecemeal approach
(Kaminski, Headey, and Bernard 2009). The approach dif-
fers substantially from the conventional approaches of
other countries, which ignored the specificities of the
institutional set-up (Jayne and others 1997). Reinforcing
the institutional framework has ensured better market
coordination along the value chain and higher levels of
contract self-enforcement, the main bottleneck to better
performance of cotton industries in the region.3 The
reforms in Burkina Faso also countermanded several gov-
ernment interventions and policies that had directly con-
tributed to inefficiency, while reallocation of activities
among the new institutions and tightened vertical coordi-
nation improved efficiency in the cotton sector.
Burkina Faso’s success in creating an efficient cotton
value chain stems from the institutional capacity for
improved contractual coordination and collective action—
something that was achieved through the creation of pro-
fessional (under free membership principles) cooperatives
of cotton growers, which substantially improved cotton
marketing and input credit repayment, yielding significant
operational cost savings. The improved institutional frame-
work has allowed cotton firms to provide better-quality
technical extension services and research in the course of
sectoral privatization and ensured coordination of the
delivery of public goods, quality control, picking and gin-
ning, and marketing activities. At the same time, access to

110 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
 Table 6.2 Reforms and Outcomes in Benin and Mali, 1992–2010
Benin Mali
Period Development Outcome Development Outcome
1992–99 Supervised liberalization of input Production increased Traditional integrated commodity Production increased
distribution; progressive by more than chain experienced financial by more than
establishment of professional and 50 percent troubles and embezzlement 50 percent
interprofessional associations
2000–05 Quasi-competition and Slight increase in Producer boycott, opening of reform Stagnation, with periodic
dismantlement of parastatal, production; dialogue; establishment of cotton declines as result of
emergence of several farmers’ problems in input cooperatives (but limited adoption farmers’ collective
networks, establishment of financial recovery de facto) action
clearinghouse, disappearance of governance
former professional associations;
institutional failures
2006–10 Move toward a private monopolistic Yields and production Reform process under way, with Production plummeted
organization with new public decreased, privatization and convergence as a result of delays in
regulation and new institutional stabilizing only toward zoning approach (splitting and uncertainties
framework for public-private recently of parastatal into four private surrounding reform;
partnership territorial entities linked through farm profitability
newly established interprofessional declined
body)

Source: Kaminski, Headey, and Bernard 2009.

finance and inputs has been eased for smallholder farmers,
and transportation services are now provided by the private
sector. Figures 6.3 and 6.4 present the main organizational
changes.
In the long term, mitigating vulnerability requires that
the main stakeholders improve their risk-management
strategies and that the government increase the quality of
agricultural services provided to farmers (research, exten-
sion, and quality grading). There is also a need to invest in
human capital in a number of areas. To deal with world mar-
ket volatility, the new price-determination mechanism—
which has suffered from lack of adjustment—has been
supplemented by an independently managed smoothing
fund. Price and weather risk-management instruments are
currently in development. Another risk-mitigation strategy
is to moderate U.S. dollar–euro exchange rate volatility
(using, for example, the holistic approach proposed by the
World Bank 2009a). Input price risk is another substantial
risk faced by cotton sector participants. In response, stake-
holders are exploring the possibility of establishing an input
fund that could lead to as much as 20 percent savings on
input procurement costs. Last, stakeholders are working to
provide a more adequate set of microfinancial and micro-
insurance instruments to farmers to stimulate small-scale
farm investment and protect against production risks,
including those related to weather. With the sustainable
management of smoothing schemes and risk-mitigation
strategies, most of the income vulnerability of cotton could
be removed.
Although the Burkinabe cotton story is largely one of
removing institutional constraints, further reforms and new
policy frameworks are needed. The growth in cotton areas
in the early 2000s reflects the lack of alternative (crop) solu-
tions, combined with a strong comparative advantage for
cotton growing in a very constrained market environment.4
Farmers need a secure access channel to output markets,
because marketing is typically too costly for them to do
on their own—or they lack the human capital to do so.
Because of market failures (incomplete or missing input
and rural credit markets), Burkinabe farmers must estab-
lish contractual arrangements with downstream stakehold-
ers (processors and traders) to access inputs and credit.5
Effective contractual arrangements involve a high degree of
market coordination and trust-building relationships. At
the same time, Burkina Faso’s challenging business environ-
ment and poor overall institutional framework discourages
entrepreneurs from investing in agrifood or large-scale
trade because of unmanageable risks and hidden costs. This
lack of investment limits the scope for developing other
high-value commodity chains, even where the potential
exists. The unsatisfactory outcome is that most tradable agri-
cultural output markets are tied to the cotton sector for
accessing inputs and credit, increasing the vulnerability of
rural income to changes in the cotton market.

CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 111
 Figure 6.3 Prereform Organizational Model of the Cotton Sector in Burkina Faso

GV Cotton producers

Input Cotton
credit seed

Other farmers
SOFITEX

Cotton fibers Cotton seed purchases,

ginning, oil mills,
purchases and distribution
of agricultural inputs, Farmers unions and faitières
Imported inputs
input credit

International cooperation (WB, AFD)

Banks DAGRIS Government

Source: Kaminski 2008.

Note: GV are village groups; AFD is Agence Française de Développement; DAGRIS is Développement des Agro-Industries du Sud (DAGRIS), a French
parastatal involved in the agroindustrial sectors of many French-speaking countries.

Figure 6.4 Postreform Organizational Structure of the Cotton Sector in Burkina Faso in 2007

Interprofessional
agreement

30% of SOFITEX
Government Cotton companies UNPCB Banks
35%

SOFITEX SOCOMA FASOCOTON Cotton Credit committees

unions
34% 100% 100%

DAGRIS Reinhardt
GPC

Research Technical agents

Cereal inputs
Management
assistance for GPCs
and cotton unions
Cotton inputs
Technical assistance
Cotton producers

Source: Kaminski 2008.

Note: SOCOMA and FASOCOTON are the regional cotton firms that were granted regional monopolies in the more marginal center and eastern cotton-
producing regions. UNPCB is Union Nationale des Producteurs de Coton du Burkina Faso; GPC is Groupement de Producteurs de Coton . DAGRIS is no longer
operating in Burkina Faso after its withdrawal from SOFITEX shareholding in 2008. The sector then went through financial restructuring with state recapital-
ization, which resulted in a substantial reduction of producers’ shares in cotton companies’ capital (from 30 percent to around 10 percent for SOFITEX).

112 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
DRIVERS OF COTTON GROWTH; MARKET within the commodity chain while the state withdrew from
AND INSTITUTIONAL ADVANTAGES most of its activities. Although most of the other key sectors
of the economy—the agricultural input supply chain, energy
Structural characteristics of rural markets partly explain
and other public utilities, and a portion of the financial
why the Burkinabe economy has remained so dependent on
sector—remained at least partially state controlled until the
its cotton sector. Historically, the cotton sector has been
early 2000s, they have been progressively privatized and lib-
better supported by institutions, stakeholders, and infra-
eralized in the years since. All banks are now privately con-
structure than cereal sectors, which face enormous logistical
trolled, as are some public utilities and input markets. This
constraints and high transport costs. The cotton sector in
completed the privatization and liberalization process that
Burkina Faso has benefited from better social organizations
occurred in the trade, transport, and cereals sectors. The evo-
of farmers, better public and private institutions, and better
lution of organizational structures in the cotton sector also
infrastructure, ensuring better-functioning markets (Bassett
has created better incentives for farmers and manufacturers
2001). Infrastructure is funded by cotton revenues through
and allowed them to benefit from profit-sharing. Apart from
private (local public goods) and public investments. Since
these improvements, stakeholders in other agricultural sec-
the 1980s investments in the cotton sector have also fostered
tors had to incur more economic risks in the course of liber-
technology adoption, thanks to high-quality research, coop-
alization (in terms of the profitability of different marketing
eration between French and local researchers, farmer organ-
channels, processing, and retailing activities), although they
izations, and top-down implementation policies.
stand to reap potentially higher financial margins.
Technological progress in agricultural sectors is more
Cotton producer organizations have benefited rural civil
likely to diffuse when land becomes increasingly scarce
society and revitalized rural communities in Burkina Faso
(Boserup 1965). Constraints in the farmer environment
and neighboring countries (SWAC Secretariat and OECD
(liquidity and market failures) hamper the adoption of
2007). Such organizations help farmers establish profes-
technology, notably for the poorest households that seek
sional structures and participate as political actors. Gover-
to meet their food-security objective and minimize pro-
nance structure and management capacities are critical to
duction risks (Abdoulaye and Lowenberg-DeBoer 2000).
this success, in addition to local social conservatism and
Fertilizer use is constrained by lack of experience, access to
other social norms at the village level. As Bernard and others
inputs and capital markets, and learning (Abdoulaye and
(2008) show, community-oriented organizations are often
Sanders 2005).6 These problems suggest the need for more
captured by traditional authorities and local elites; in con-
extension services, quality control, and research to foster
trast, when elaborate administrative rules are established,
fertilizer use on higher-value crops and efforts to reduce
market-oriented organizations are much more democra-
the risk associated with fertilizer use through training and
tized and efficient. In the cotton sector, market-based
quality standards for lower-value crops. As a result of
cooperatives of cotton farmers have been established and
demographic pressure in and migration into cotton areas,
producers have become significant players in policy making
as well as the relatively high value of cotton compared to
and the public debate. In other rural sectors, by compari-
other crops, cotton has attracted increasing levels of tech-
son, institutional constraints remain more or less unchanged
nology, capital, and inputs.
at the farmer level.
The advantages of growing cotton (rather than other
Importantly, cotton provides better risk-management
crops) include better access to rural finance and agricul-
arrangements to producers than do other sectors, especially
tural inputs, thanks to interlinked transactions with
with regard to price, outlets, and marketing. Cotton pro-
powerful agribusinesses (contract farming and other con-
ducer prices are administratively determined at the begin-
tractual arrangements) and the ability to overcome inher-
ning of the cropping season and the pricing formula allows
ent market failures in the realms of input and rural credit
for ristournes (after-harvest bonus payments) to farmers. In
markets. The strong position of these agribusinesses
addition, cotton farmers have better access to market infor-
ensures contract self-enforcement (in that it prevents sig-
mation than farmers in other sectors, thanks to their effec-
nificant room for side-selling). This is not the case in the
tive organizations and national union, expansion of local
cereal sector, which is characterized by small traders who
infrastructure, and increased use of mobile phones in
lack physical capital.
cotton-producing areas. Better information sharing can also
In Burkina Faso, most reforms in the cotton sector
help agribusiness entrepreneurs provide insurance schemes
addressed government failures and institutional weaknesses by
to farmers.
building the capacity of producers and upgrading institutions

CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 113
 Despite several shortcomings (including the weakness of
the farmers’ union, corruption, and adjustment lags in the
price determination mechanism), the revamped cotton
policy framework in Burkina Faso has contributed to an
increased concentration of cotton exports and dependency
on cotton production at both the national and farm levels.
Consensus building was forged by top-down establishment
of the cotton union, state authoritarianism, trust-building
with foreign partners, and accumulation of social capital at
the village level (Kaminski and Bambio 2009). Consensus
building followed by institutional innovations built part-
nerships along the supply chain with the professionalizing
of farmers’ groups and reinforced their bargaining power.
As a result of upgraded institutional and policy frame-
works, profit-sharing has evolved in favor of producers,
while the management, input procurement and cost recov-
ery, and overall market coordination of the sector has
become more efficient, even in the face of the decline in the
world price of cotton and the appreciation of the euro
against the U.S. dollar. A larger profit share for farmers
reinforced production incentives once coordination was
improved, contributing to the impressive cotton growth
pattern observed in the 2000s (figure 6.5).
Aside from these positive outcomes, the state bailout of
SOFITEX and the lack of political sustainability of the
cotton policy framework have reduced sectoral efficiency
and led to failures in market coordination once the world
Figure 6.5 Farm Gate and World Price Indexes of Seed Cotton, 1990–2009
14
12
10
0
90
92
94
96
98
19
19
19
19
19
20
Burkina Faso farm gate price index, 10 = 2000
Source: Author, based on monthly data on international prices and exchange rates from data construction from IMF 2010 and Burkina Faso
national sources.
Note: The Liverpool cotton price index is the index of world cotton prices. All series were calculated from averages of monthly data covering the period
February–January, the production year for Burkina Faso’s cotton sector.
114 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
price fell in the late 2000s. As a result, smallholders’ prof-
itability is threatened and trust in the newly established
arrangements and partnerships declined from 2007 to
2009. Indeed, due to a misapplication of the pricing for-
mula in 2006–07, the smoothing fund was depleted and
farm gate prices were adjusted downward in 2007–08
once the formula started to be rigorously applied. In addi-
tion to growing deficits, SOFITEX had trouble paying
farmers and delivering inputs to them on time. Some
observers point to management problems in SOFITEX
once DAGRIS left the shareholding and the company was
recapitalized by the state. But the recent spike in cotton
prices, even in the face of higher marketing costs attribut-
able to the political crisis in Côte d’Ivoire, means that
financial prospects are optimistic in the mid-run. The
expectation of good financial prospects is all the more
true since the price-determination mechanism has been
in effect (and rigorously applied) over the past five produc-
tion cycles. Burkinabe farmers’ profitability is tightly linked
to world prices—with predefined pricing at the beginning
of the crop season, they receive roughly 60 percent of the
world price of cotton for their crops.
In sum, while growth strategies aimed at economic diver-
sification should be pursued within an adequate policy
framework, governance, management, and institutional
improvements are still required in the cotton sector if it is to
remain a sustainable pillar of the Burkinabe economy.
00
02
04
06
08
20
20
20
20
Liverpool price index, 10 = 2000
COMMODITY DEPENDENCE AND ECONOMIC tend to vanish, and ensuring a minimum amount of agri-
VULNERABILITY cultural inputs to the noncotton agricultural sectors.
The cotton-led growth of Burkina Faso has been nega-
Over the past decade, GDP growth in West African coun-
tively affected by volatile world cotton prices (competing
tries has been derived largely from agriculture (export crops
synthetic fibers, cotton subsidies in the United States and
in coastal countries and livestock in landlocked countries)
Europe) and the appreciation of the euro (World Bank
and mining. As many economists, including Hausman,
2009a).7 Moreover, increases in production have been driven
Hwang, and Rodrik (2007), note, export sectors are key to
mainly by increases in factors of production (labor, land,
economic growth, as they typically have a cascade of effects
inputs) rather than through increases in total factor produc-
on other sectors. However, export earnings cannot compen-
tivity. Economic vulnerability from commodity dependence
sate for weak internal conditions for sustained growth.
has thus been exacerbated by the nonintensive pattern of
The problematic issue here is commodity dependence.
growth, which has had a limited effect on overall economic
Because of few absolute advantages, poor countries such as
transformation and limited spillovers through domestic
Burkina Faso often experience unbalanced growth pat-
demand-led growth. Due to environmental and demo-
terns driven by a restricted number of export commodities
graphic limitations, this growth path looks to be unsustain-
traded on very volatile world markets. Furthermore, these
able in the long run. In general, the Burkinabe economy is
commodity markets are characterized by distortive poli-
still highly agriculture-oriented with a growing importance
cies and barriers to entry or participation for farmers in
of cotton during the late 2000s and increasing importance of
developed countries. In addition, production is subject
the mining sector. Figure 6.6 compares the economic growth
to climatic variability (droughts and floods) and prob-
of each economic sector over the past decade.
lems of soil fertility. Relying on a few commodities with
Economic diversification is one of Burkina Faso’s
uncertain profitability leads to economic vulnerability
national priorities, as stressed in recent official strategic
and calls for an in-depth examination of viable economic
documents. Of particular importance is diversification in
alternatives.
the rural sector toward higher-value agricultural exports,
Globally, the outlook for cotton profitability is uncertain
such as fruits and vegetables. In the medium term, diversifi-
because of enduring institutional and governance limita-
cation to manufactured products does not appear necessary
tions (despite recent improvements in both areas) and
for growth. In contrast, increases in the production of
volatile world markets. In addition, cotton sectors some-
higher-value natural-resource-based products are required.
times still serve political ends—for example, offering good
Such a shift may raise income levels by developing capacities
prices to producers before elections, even if these offers

Figure 6.6 Real GDP Average Growth, by Economic Sector, 1994–2008

16

14

12

10
Percent

0
re

w and

om d

es

s
ck

ice
in

rin

ad
io

lec n

vic
tu

to

te rt a
in

ct

Tr

rv
tu

e
ul

,
s

as

er
ru
at
ve

se
ac
ric

o
,g

ls
st

sp
Li

uf
Ag

er
on
ty

cia
an
an

th
ici

an
Tr
M

O
tr

Fin
ec
El

Source: World Bank, CEM Burkina Faso, and IMF 2009.

CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 115
 for manufactured goods together with the capacities of
marketing and exporting a diversified basket of primary
products. The application of technology to increase cotton
profitability should also be encouraged.
Although growth of the cotton sector has been accom-
panied by some amount of economic diversification, that
diversification has not been sufficient to stimulate struc-
tural transformation through macroeconomic spillovers.
Although cotton provides skills to the other agricultural
sectors, the growth of the cotton sector has not brought
improvement in labor and land productivity and has not
enabled sufficient redistribution of public goods such as
education and health. In addition, spillovers of cotton
activities to other economic sectors in Burkina Faso are
restricted because of the limited industrial nature of cotton
growing and its limited size in the rural sector (not more
than 20 percent). A natural way to encourage spillovers
would be to scale up textile activities. Finally, it is impor-
tant to recognize that the increase in demand for agricul-
tural inputs from cotton growers themselves has not been
met by corresponding increases in local supply. Most agri-
cultural inputs and nonfood goods are produced abroad,
limiting transmission of demand increases in the rest of the
economy.
Constraints to economic diversification
A number of factors constrain the diversification of Burkina
Faso’s economy.8 Inefficient and incomplete policy reforms,
institutions, and investments (in infrastructure, capacity-
building, and public goods, for example) are all hindrances
to diversification. Yet diversification is clearly a necessity for
more sustainable and balanced growth. Macroeconometric
studies (Berthelemy 2005) show that diversification indica-
tors positively influence GDP growth rates, notably
through increases in total factor productivity.
In general, there is a U-shaped relationship between eco-
nomic concentration in exports (the Herfindahl index) and
per capita income (World Bank 2009a). In the low part of
the curve (where the relationship is negative) stand low-
income countries such as Burkina Faso, characterized by
export earnings derived mainly from agriculture (Imbs and
Wacziarg 2003). The key role of diversification is to smooth
export earnings derived from volatile (agricultural and
resources) world markets in developing countries (Massel
1970), which fosters economic structural changes and con-
solidation. A more sustainable and balanced growth path
can be attained by economic diversification and a set of
policies and institutions that accompanies the process of
116 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
diversification by establishing facilitating conditions. This
process involves general policies aimed at building insti-
tutions, providing necessary public goods, and loosening
economic constraints in areas such as public administra-
tion, market imperfections and transaction costs, fiscal and
trade policies, and weak governance. Such policies should be
combined with sectoral policies according to absolute and
comparative identified advantages, emerging champions
(World Bank 2009a), and the structural role of key eco-
nomic sectors (Hirschman 1958). Hence, diversification and
specialization may exist side by side because of segmenta-
tion of the economy. This is the case for most West African
countries, where there is a lack of economic integration
between the primary sector (agriculture and mining) and
the industrial sector.
The most important challenge for spurring sustainable
growth in Burkina Faso lies in increasing total factor pro-
ductivity, which depends to a large extent on the investment
climate. Policies that make it easier to do business, reduce
the burden borne by entrepreneurs, and upgrade the insti-
tutional framework are worth considering. From an invest-
ment climate perspective, Burkina Faso has improved
significantly over the past decade, placing it in the
midrange of African countries in the World Bank’s Doing
Business rankings (World Bank 2009b) and other indexes.
Yet important bottlenecks remain, such as lack of access to
finance and electricity, high tax rates, unfair (on fiscal
grounds) practices in the informal sector, corruption,
inadequate transportation, lack of access to land, ineffec-
tive bureaucracy, lack of technology, inadequate health
services, and lack of an educated labor force. These con-
straints raise labor costs, limiting producers’ competitive-
ness. Large firms in Burkina Faso, however, are roughly as
productive as smaller ones (World Bank 2009a), implying
that differences in labor productivity are not necessarily a
matter of scale efficiency but rather rely on higher capital
intensity for larger firms. Indirect and “invisible” costs
account for an estimated 8.2 percent of firms’ sales in
Burkina Faso (table 6.3).
Constraints to technology adoption and diffusion
Promising technologies exist in the rural sector, but low
levels of technology adoption and diffusion hamper the
process of diversification.9 Several underlying constraints
are at play. First, diffusion of a new technique or technol-
ogy requires complementary techniques and inputs, for
which extension services and sustainable arrangements to
ensure the provision of inputs are needed. Given the market
 Table 6.3 Indirect and Invisible Costs as Share of Firms’ Sales, Selected Sub-Saharan African Countries
(in percent)
Invisible costs
Cost of
Indirect costs Cost of corruption Total
Losses manager’s (informal indirect
due to time to payments to Cost of and
power deal with get things security invisible
Country/year Transport Telecommunications Customs outages regulations done) measures costs
Burkina Faso 2006 2.23 0.70 1.63 1.20 0.05 1.48 0.86 8.20
Benin 2004 — — 0.48 1.16 0.06 4.27 0.57 6.55
Mali 2007 1.96 0.86 0.26 1.39 0.05 0.24 0.12 4.60
Uganda 2006 1.08 0.50 0.52 28.16 0.06 2.18 0.43 32.93
Zambia 2007 0.61 0.71 0.68 1.79 0.04 0.05 0.93 4.81

Source: World Bank Enterprise Surveys, 2004–07.

Notes: Estimates are based on at least 15 observations. — = not available.

failures affecting rural credit and agricultural inputs in
Burkina Faso, it is understandable that technology diffu-
sion may be constrained.
Apart from institutional and market constraints, technol-
ogy diffusion and adoption need a clear and consistent pol-
icy framework for new marketing strategies, risk-mitigation
strategies, or export promotion set-ups to be sustainable
and provide the necessary incentives for long-term invest-
ment. In the rural sector, Abdoulaye and Sanders (2006)
have shown that new marketing strategies (such as the
Intsormil project10) do not ensure sufficient investment
incentives if food security policies are inconsistent and do
not allow a greater number of farmers to derive more prof-
its during adverse years (removing export quotas, for
instance). Technology adoption also relies on comprehen-
sive exchange rate and trade reforms. In the manufacturing
and services sectors, policy frameworks are also critical. The
needed reforms combined with long-term commitment of
the public sector in export promotion and investment
strategies should, in turn, facilitate technology diffusion and
diversification in higher-value production and exports.
Market constraints: Missing and incomplete markets
in the rural sector
Agricultural producers face missing and incomplete mar-
kets for their goods. This is especially true for cereal pro-
ducers. Cereal markets are still characterized by high
transaction costs, ineffective institutions, and discrimina-
tory policies, maybe because of their low value to weight
ratio and logistical disadvantages (information access,
inland freight cost). This notably contrasts with the dynam-
ics of cotton markets. Hence, the transport cost component
of cereal profit margins is significant, meaning variability at
producers’ gate prices when combined with low market
integration. Public investments in market infrastructures
have been low, and only some rural cotton producing areas
have been unlocked thanks to cotton benefits and the inter-
ests of agribusinesses. Not surprisingly, the most integrated
cereal markets are the ones located close to the most pro-
ductive cotton markets, because they benefit from better
infrastructures and institutions.
After the liberalization of cereal markets in Burkina Faso,
many small-scale traders coexisting with larger private
traders have replaced government marketing. However,
because of information asymmetries with farmers and the
lack of access to reliable information services, traders have a
comparative advantage and better information regarding
local and central markets, retailers, and processing facilities.
Even if this has created competitive emulation, there has
been an associated lack of investment in human capital and
innovation, few long-term contractual relationships with
agribusinesses, and no significant change in access to finan-
cial markets. Due to several market failures and institutional
weaknesses (contract enforcement), informal traders rely on
social networks, rendering competition imperfect (Barrett
1997). In addition, Burkina Faso’s network economy—
retention of profits among insiders—tends to limit business
expansion, as noted by Badiane and others (1997).
Not only do producers face constraints in output mar-
kets, which slows down diversification, but they also face
constraints with respect to input markets. Hence, the poten-
tial for income-enhancing diversification of the rural sector
is limited. Input market imperfections result from struc-
tural deficiencies such as high transaction costs, liquidity
constraints of producers, and asymmetric information.

CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 117
 Lack of well-maintained and high-quality infrastructure
also impedes market integration in Burkina Faso, translat-
ing into higher market risks borne by producers and traders,
together with capacity constraints (risk-management and
storage).11 Cost-effective ways to provide infrastructure
may use existing infrastructures with the involvement of
user communities (mainly for maintenance with own-
managed funds). Productivity benefits could be achieved by
improving off-road transport and intermediate means of
transport with capital-savings techniques for road construc-
tion or using labor-based techniques to overcome usual
problems related to equipment use and availability in the
region. Soft infrastructure, such as information, is also in
short supply. Diversification from cotton to other agricul-
tural commodities needs to rely on better market informa-
tion services and farmer literacy to foster adoption of more
profitable crops and techniques.
Institutional constraints
A country’s institutional environment is composed of
market and nonmarket relationships between different
stakeholders. The role of institutions is all the more cru-
cial in Burkina Faso given that markets are subject to
structural deficiencies and constraints for participants.
Nonmarket institutions define social rules and norms
that, in turn, overcome the other market and institutional
imperfections. But informal institutions, such as social
networks and personal relationships, have been shown to
be less efficient than formal ones, and they severely
restrict the degree of competition, information, and busi-
ness expansion.
Supply chains and vertical relationships (regula-
tory, informal institutions, contract enforcement,
information asymmetries). In the formal realm, the
institutional structure of many supply chains was deeply
reconfigured in the course of structural adjustment plans
and sectoral reforms. Vertical relationships now entail
more specific arrangements between stakeholders and
more coordinated provision of public goods. Stakehold-
ers face more options as marketing channels may be
potentially beneficial for them if stakeholders can
improve their capacities (information, management, bar-
gaining, storage, infrastructures). However, competition
in several sectors can be detrimental because of the per-
sistence of other market failures (Lipsey and Lancaster
1956) and the fact that it can lead to coordination prob-
lems regarding the provision of certain public goods,
118 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
such as agricultural or industrial innovation and quality
grading.
In Burkina Faso organizational structures rely on formal
and informal institutions to coordinate the decisions of
major players and to regulate the industry. Formal structures
include state rules and formal regulations or collective organ-
izations such as producers’ associations, partnerships, and
farmer unions. The evolution of the regulatory framework in
Burkina Faso can be qualified as uneven. Overregulation in
several sectors has been quite counterproductive and has
notably contributed to restrictive environments, such as in
the trade sector. Despite several improvements, there is a per-
sistent lack of confidence in formal institutions, including the
legal system. There are no effective institutions dedicated to
investment and export promotion, indicating the absence of
a long-term strategic vision for economic development and
diversification. In contrast, standardization and quality pro-
motion systems are being developed, and business associa-
tions are beginning to adopt quality control systems within
their marketing systems. Continued improvements in these
areas would benefit the whole set of private investors and
business expansion.
The institutional environment. Because formal insti-
tutions in Burkina Faso face a lack of credibility, many
producers and private stakeholders rely on informal insti-
tutions, social networks, and personal relationships. How-
ever, this poses a problem, since the prevalence of informal
institutions weakens the credence of the formal ones.
Informal village norms—such as those for credit, risk, or
access to land—also enable local actors to overcome several
market constraints. In matters of land access, however,
informal norms often entail exclusion of some ethnic
groups or difficult access to the fertile or accessible land for
some societal groups. Political and ethnic biases in the allo-
cation of land imply allocation inefficiencies. Informal and
traditional risk-sharing arrangements provide only imper-
fect assurance of low effectiveness.
The decentralization of rural development policies in
Burkina Faso, together with more participatory approaches,
has contributed to a democratization process whereby
responsibilities have been gradually transferred from
national institutions to newly established local institutions.
However, capacity constraints are still very strong.
To improve the business environment, national stake-
holders in Burkina Faso have begun exploring various
options: public-private partnerships, business associations
and interprofessional bodies, information services, and
consensus-building institutions. While all of these options
have the potential to improve vertical and contractual rela-
tionships within supply and value chains, the experience of
building the interprofessional association in the cotton sec-
tor was especially encouraging in that it provided a credible
institutional framework to the main stakeholders and
ensured market coordination while being limited by the
predominant position of SOFITEX, which can sometimes
impose some decisions on collective management issues
(quality grading, input distribution system or regulation,
and so on).
Risk management. Along with being a market constraint,
risk is an institutional constraint. Inefficiencies in the cur-
rent institutional arrangements in Burkina Faso require
mitigation of business and production risks as well as eco-
nomic ones. For example, minimizing production risks for
all agricultural commodities and livestock requires a better
institutional framework. The development of new
microinsurance schemes may help farmers facing external
shocks to production. Indeed, improved transmission of
information about weather conditions would help in the
development of weather index microinsurance schemes
(as already achieved in Southern Africa), providing farmers
with producers’ incentives. Another innovative insurance
scheme that could be applicable is parametric insurance.
Although the development of markets in innovative insur-
ance schemes requires a strong financial framework to be
established, enhancing information access through new
technologies for information and communication could
help spread these schemes.
Political constraints
Constraints to economic diversification also include the
ones that come from policy making and that contribute
to market and institutional failures. Most of them can
be classified as government failures: public governance
problems, transparency, democratization, nonbenevolence,
corruption, and lack of foresight. Despite substantial
improvements in the political environment in Burkina Faso
over the past decade, enduring political practices hamper
the effectiveness of upgraded policies and the questions of
democratization and authoritarianism remain.
Commodity reforms and the political economy of
cotton-cereal reforms. Although commodity reforms
in many Sub-Saharan African countries have been the
subject of criticisms (Akiyama and others 2001), Burkina
Faso has been viewed as a positive experience. These suc-
CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
cesses can be attributed to a consistent policy framework
in which the state gained credibility with its economic
partners, farmers’ groups were strengthened, and private
arrangements (such as contract farming) were not hin-
dered by second-generation controls. Policy inconsistency
remains in the cereal sector, however, in the form of food
security policies (lowering agricultural profits of the most
efficient farmers), protectionist trade policies (lowering
incentives for diversified exports), and restrictions on
input markets. Fortunately, these short-lived interven-
tions that were common in the past are increasingly less
evident. If needed, they should be decided upon harmo-
niously at the regional level, one of the purposes of the
Comprehensive Africa Agriculture Development Pro-
gramme (CAADP), an initiative of the New Partnership
for Africa’s Development (NEPAD) that calls for more
harmonized agricultural and trade policies at the regional
level in West Africa.
Sufficient agricultural extension services remain a chal-
lenge in Burkina Faso. Currently, there is a lack of means,
training, and funding for extension services, along with lack
of involvement by the state.
Political economy, social contracts, and the urban
bias. The political economy of agricultural policies
explains the persistence of contradictory policies and
incomplete market reforms. Compared to neighboring
countries, Burkina Faso stands between the situation of
Mali—steady adherence and commitment to market
reforms—and that of Benin—de jure reforms and de facto
state control with regard to noncotton sectoral policies (the
reverse with regard to cotton policies). Private incentives in
Burkina Faso have been jeopardized by entrepreneurs’ fear
of government intervention and by the somewhat volatile
political environment. Fewer private sector opportunities
create scope for a larger government involvement in mar-
kets, which, in turn, fuels this pattern. Therefore, political
economy conditions have contributed to constraining agri-
cultural diversification while favoring the cotton sector.
These political conditions also apply to the services and
industrial sectors in Burkina Faso.
The political environment. Despite several improve-
ments, the political environment in Burkina Faso contin-
ues to present constraints in the form of enduring
political practices resulting in weak or poor governance,
collusion of interests between officials and executives, and
lack of political and democratic accountability of political
leaders.
119
 CONCLUSION constraints described in this study in Burkina Faso are still
of crucial importance. This calls for a policy-led economic
From an overarching perspective, cotton continues to
transformation induced by structural changes, a consistent
account for a significant share of exports in Burkina Faso’s
policy framework, and institutional innovations. Innova-
very concentrated economy. Future global demand for cot-
tion needs to take place in a range of sectors: agriculture and
ton, however, does not appear consistent with an economic
livestock, mining, trade, tourism, and the food industry.
strategy in Burkina Faso that relies heavily on cotton
In the mining sector, the Burkinabe government has
exports as a driver for growth, not least because world cot-
recently established a mining law to improve incentives for
ton demand is expected to remain at volatile levels because
cost-effectiveness, export promotion, and for informal
of imperfect substitution among textile fibers and rising oil
orpailleurs. In the agricultural sector, income-enhancing
prices (oil is used in the production of synthetic fibers). But
options exist and are achievable but would not lead to sig-
in the short run, there is a need to maximize cotton rev-
nificant changes in the current environment of smallholders
enues and benefit from the historic price swings on the
and traders, partly because of enduring market imperfec-
world markets.
tions, structural deficiencies, and macroeconomic instabili-
A useful approach for Burkina Faso would be to exam-
ties. Higher-value production is also of potential interest
ine other sectors with significant export growth potential,
and has already driven growth and diversification in India
such as gold and livestock. Ramping up exports in these
and in several countries in Eastern Africa (for example, in
sectors may reduce dependence on cotton incomes.
horticulture, fruits, and nuts). A strategic set of sequenced
Expanded production and export of other agricultural
actions could be led by officials and local leaders.12
products (vegetables, fruits, poultry, and rice) and nonpri-
Another area in which Burkina Faso has space for
mary commodities may also serve to reduce the economic
reform is trade. Currently, the trade environment in the
concentration on cotton.
country is characterized by high tariffs and low market
Even with the potential expansion of other sectors, cot-
integration. Burkina Faso would also benefit from diversi-
ton is likely to continue providing livelihoods for a signif-
fying its primary products, the production of which is con-
icant share of Burkina Faso’s rural population and to
strained by the scope of development of manufacturing
remain important in the national economy. The most
industries (for example, the food industry). Reasons typi-
urgent concern is to tackle the issue of “sustainable inten-
cally invoked for the limited scope of manufacturing in the
sification” at the farm gate, together with increasing the
country include its abundance of land, scarcity of skilled
cost-effectiveness of cotton firms. Overall, there is a need
labor, insufficient or poor-quality infrastructure, lack of
to consolidate the cotton sector in the course of ongoing
policy reform, high transaction costs, and mismanaged
reforms and institution building, and to initiate a signifi-
or uninsured risks (Collier 1998, 2002; Habiyaremye and
cant process of economic diversification through the for-
Ziesemer 2006; Collier and Gunning 1999). A consistent
mulation and implementation of growth-enhancing
policy framework toward structural transformation should
policies aiming to overcome the most relevant con-
deal with these constraints.
straints. Nonmarket factors, such as the quality and effec-
Several of the issues raised in this study are already being
tiveness of institutions, are also at play, meaning that
addressed by officials and foreign donors in policy formula-
improvements in the business environment and invest-
tion and project implementation. Stylized facts, in fact,
ment climate would not automatically lead to a scale-up of
show that policy makers have been successful in bringing
nontraditional export industries as a prerequisite for
about change in several areas:
structural transformation of the economy.
While policies are moving in the right direction—
generalization of GM Bt cotton (genetically modified cot- ■ Established priorities: food security, poverty reduction,
ton with Bacillus thuringiensis inserted into the seeds) and private investments, and public-private partnerships in
diffusion of improved agricultural productivity practices infrastructure and human capital.
(such as irrigation) and soil conservation, improved busi- ■ Policies for emerging and developing economic sectors
ness environment, implementation of more adequate risk- ■ On-going reforms
management instruments and the new accelerated growth ■ Strategic action plans with donors: special economic
strategy, explicit support to growth poles as a means to zones and planning for employment, competitiveness,
create faster growth and diversification—removing the and diversification of economic growth poles

120 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
■ Capacity to improve the institutional framework (for
example, the cotton reform) even in the face of external
shocks (declining world cotton prices and economic and
political crisis in neighboring Côte d’Ivoire)
■ Accumulated experience
■ Government capacity to react to shocks and resistance to
economic turmoil while maintaining the pace of reforms.
The outlook for Burkina Faso—albeit promising—could
be made more secure with scaled-up investments in crucial
sectors, and deepening market and political reforms. As
such, the road is half-traveled. Burkina Faso is still at the
stage of establishing policy and institutional frameworks.
Governance has mixed records, despite progress in areas
such as public investments. Demographic growth is a key
challenge but the way government is responding to it, rather
than hiding behind cultural values and norms, may be
viewed as a success. Creating and using fiscal space in
the near future, coping with demographic growth, and pur-
suing reforms will all be crucial. Key investments in infra-
structure (first priority) and in human development (second
priority) will be the cornerstone of subsequent development.
While progress thus far has been in the right direction, it
must be complemented by sector-specific policies in line
with structural economic change and better promotion of
private investment and exports. Several further improve-
ments in the public administration and the political systems
are also expected. Indeed, internal political sustainability of
the policymaking and decision processes is a necessary con-
dition for successful policy-led development and reforms
(Rodrik 1996).
Finally, in the course of market reforms in Burkina Faso,
policy makers should bear in mind that removing market
imperfections (with investments in infrastructure to reduce
transaction costs) needs to be done very carefully. As Lipsey
and Lancaster (1956) note, the persistence of other market
failures may have an adverse effect even after imperfections
are removed. Timing of market and institutional reforms
and infrastructure investment should proceed accordingly.
NOTES
1. Other significant sources of export earnings are live-
stock and gold.
2. The higher incidence of poverty in rural areas is dis-
putable, because the official poverty line established by the
Institut National de la Statistique et de la Démographie
(INSD) in 2003 places too much weight on the prices of
necessary goods (such as staples and water) relative to the
CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
weights used to construct the urban CPI (Grimm and
Gunther 2004).
3. Market coordination is essential to ensure the delivery
of crucial public goods (inputs and credit) in an incomplete
marker environment. More competition broadens the scope
for strategic defaulting on input credit and side-selling
(selling outside the contractual arrangement to another
company) and reduces the incentives to provide public
goods, such as research or extension services.
4. This phenomenon explains the rise in the Herfindahl
index from 0.2 to 0.6.
5. Cotton inputs are provided by ginning companies,
which also directly purchase and gin seed cotton; inputs
have been subsidized lately by the government.
6. The use of fertilizer increases with roads and rainfall.
7. Cotton production still ensures most cash earnings for
farmers as well as most of their inputs and credit for other
crop productions. Cotton earnings also provide funds for
increasing capacities of smallholders’ unions, agricultural
extension services and research. The strong linkages
between cotton and other crop productions must be borne
in mind while not forgetting agronomic and economic
complementarities that are harnessed in a very constrained
environment for farmers.
8. Economic diversification is defined here as the process of
progressive enlargement of the number of goods produced
in an economy without a decline in productivity.
9. Promising technologies include integrated livestock-
crop farming systems; soil conservation; regreening pro-
grams; small-irrigation schemes; intercropping and alley
cropping; pest management; early cultivars for rain-fed rice,
millet, and sorghum; and new generations of maize varieties
(Kaminski 2008).
10. They involve the widespread use of inventory credit,
agroprocessing of traditional cereals such as millet, producers-
processors contracts in the poultry industry. and so on.
11. As noted by Platteau (1996), infrastructural constraints
are a major cause of the low long-run supply response of
farmers to price incentives, notably for transport and com-
munications. Although tax reforms and infrastructure
investment both compete for public funds, their effects are
complementary and differently impact producers.
12. The mitigation of agricultural incomes’ vulnerability to
weather and external shocks is a key strategy through
income-enhancing diversification under intensified farming
systems for selected products (comparative advantages and
market/trade potential), an improved business environment
to encourage private sector investment (decreasing trans-
action, financial, labor, and energy costs), and an improved
trade environment (infrastructures, export promotion,
market integration, and quality standards).
121
 BIBLIOGRAPHY
Abdoulaye, T., and J. Lowenberg-DeBoer. 2000. “Intensifica-
tion of Sahelian Farming Systems: Evidence from Niger.”
Agricultural Systems 64 (2): 67–81.
Abdoulaye, T., and J. H. Sanders, 2005. “Stages and Determi-
nants of Fertilizer Use in Semiarid African Agriculture: The
Niger Experience.” Agricultural Economics 32: 167–79.
———. 2006. “New Technologies, Marketing Strategies and
Public Food Policy for Traditional Food Crops: Millet in
Niger.” Agricultural Systems 90: 279–92.
African Economic Outlook. n.d. “Burkina Faso.” http://www
.africaneconomicoutlook.org/en/countries/west-africa/
burkina-faso/.
Akiyama, T., J. Baffes, D. Larson, and P. Varangis. 2001.
Commodity Markets Reforms: Lessons of Two Decades.
Washington, DC: World Bank.
Badiane, O., F. Goletti, M. Kherallah, P. Berry, K. Govindan,
P. Gruhn, and M. Mendoza. 1997. Agricultural Input and
Output Marketing Reforms in African Countries: Final
Report. Washington, DC: International Food Policy
Research Institute.
Barrett, C. B. 1997. “Food Marketing Liberalization and
Trader Entry: Evidence from Madagascar.” World Devel-
opment 25 (5): 763–77.
Bassett, T. 2001. The Peasant Cotton Revolution in West
Africa, Côte d’Ivoire, 1880–1995. Cambridge: Cambridge
University Press.
Bernard, T., M-H. Collion, A. de Janvry, P. Rondot, and
E. Sadoulet. 2008. “Do Village Organizations Make a
Difference in African Rural Development? A Study for
Senegal and Burkina Faso.” World Development 36 (11):
2188–204.
Berthelemy J-C. 2005. “Commerce international et diversifi-
cation économique. “ Revue d’Economie Politique 115 (5):
591–611.
Boserup, E. 1965. The Conditions of Agricultural Growth: The
Economics of Agrarian Change under Population Pressure.
Chicago: Aldine.
Boughton, D., and T. Reardon. 1997. “Will Promotion of
Coarse Grain Processing Turn the Tide for Traditional
Cereals in the Sahel? Recent Empirical Evidence from
Mali.” Food Policy 22: 307–16.
Collier, P. 1998. “Globalization: Implications for Africa.” In
Trade Reform and Regional Integration in Africa, ed.
Z. Iqbal and M. S. Khan. IMF Institute, International
Monetary Fund, Washington, DC.
———. 2002. Primary Commodity Dependence and Africa’s
Future. Washington, DC: World Bank.
Collier, P., and J. W. Gunning. 1999. “Why Has Africa Grown
Slowly?” Journal of Economic Perspectives 13 (3): 3–22.
122 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
Conley, T. G., and C. R. Udry. 2000. “Learning about a New
Technology: Pineapple in Ghana.” Yale Economic
Growth Center Working Paper 817, New Haven, CT.
Coulter, J., A. Goodland, A. Tallontire, and R. Stringfellow.
1999. Marrying Farmer Cooperation and Contract Farm-
ing for Agricultural Service Provision in a Liberalizing
SSA. ODI Natural Resources Perspectives, 48, Overseas
Development Institute, London.
Fafchamps, M. 2004. Market Institutions in Sub-Saharan
Africa: Theory and Evidence. Cambridge, MA: MIT Press.
Foster, A. D., and M. R. Rosenzweig. 1995. “Learning by
Doing and Learning from Others: Human Capital and
Technical Change in Agriculture.” Journal of Political
Economy 103 (6): 1176–1209.
Gray, Leslie C., and Michael Kevane. 2001. “Evolving Tenure
Rights and Agricultural Intensification in Southwestern
Burkina Faso.” World Development 29 (4): 573–87.
Grimm, M., and I. Gunther. 2004. “How to Achieve Pro-
poor Growth in a Poor Economy? The Case of Burkina
Faso.” Working paper, University of Göttingen, Germany.
Habiyaremye, A., and T. Ziesemer, 2006. “Absorptive Capac-
ity and Export Diversification in Sub-Saharan African
Countries.” UNU-MERIT Working Paper Series. United
Nations University, Maastricht Economic and Social
Research and Training Centre on Innovation and Tech-
nology, Maastricht, the Netherlands.
Hausmann, R., J. Hwang, and D. Rodrik. 2007. “What
You Export Matters.” Journal of Economic Growth 12 (1):
1–25.
Hausmann, R., and B. Klinger. 2006. “Structural Transforma-
tion and Patterns of Comparative Advantage in the Prod-
uct Space.” CID Working Paper, Center for International
Development, Harvard University, Cambridge, MA.
Hirschman, A. 1958. The Strategy of Economic Development.
New Haven, CT: Yale University Press.
Imbs, J., and R. Wacziarg. 2003. “Stages of Diversification.”
American Economic Review 93 (1): 63–86.
IMF (International Monetary Fund). 2007. Burkina Faso.
Propositions de réformes fiscales: Simplification, équité et
efficacité. Washington, DC.
———. 2008. Burkina Faso: Third Review Under the Three-
Year Arrangement under the Poverty Reduction and
Growth Facility. Washington, DC.
———. 2009. Burkina Faso: “Third Review under the
Three-Year Arrangement under the Poverty Reduction
and Growth Facility. Staff Report.” Press Release on the
Executive Board Discussion. IMF Country Report No.
09/38, Washington, DC (January).
Jayne, T., and S. Jones. 1997. “Food Marketing and Pricing
Policy in Eastern and Southern Africa: A Survey.” World
Development 25: 1505–27.
Jayne, T., J. D. Shaffer, J. M. Staatz, and T. Reardon. 1997.
“Improving the Impact of Market Reform on Agricul-
tural Productivity in Africa: How Institutional Design
Makes a Difference.” MSU International Development
Working Paper 66, Michigan State University, East
Lansing, MI.
Jones, S. 1998. Liberalised Food Marketing In Developing Coun-
tries: Key Policy Problems. Oxford Policy Management.
Kaminski, Jonathan. 2007. “Interlinked Agreements and the
Institutional Reform in the Cotton Sector of Burkina
Faso.” Working paper, Atelier de Recherche Quantitative
Appliquée au Développement Économique, Toulouse
School of Economics, Toulouse, France.
———. 2008. “Cotton-Cereal Systems in West and Central
Africa: Opportunities and Constraints for Revenue-
Raising Diversification and Marketing Strategies.” FAO
All-ACP Background Paper, European Union, Rome.
Kaminski, J., and Y. Bambio. 2009. “The Cotton Puzzle in
Burkina Faso: Local Realities versus Official Statements.”
Paper presented at the annual meeting of the African
Studies Association, New Orleans, LA.
Kaminski, J., D. Headey, and T. Bernard. 2009. “Navigating
through Reforms: Cotton in Burkina Faso.” In Millions
Fed: Proven Successes in Agricultural Development.
Washington, DC: International Food Policy Research
Institute.
Kaminski, J., and A. Thomas. 2009. “Land Use, Production
Growth, and the Institutional Environment of Smallhold-
ers: Evidence from Burkinabe Farmers.” CAER-LERNA
Working Paper, Hebrew University, Jerusalem, and Insti-
tut National de la Recherche Agronomique, Toulouse
School of Economics, Toulouse France.
Labaste, Patrick. 2009. Background paper for the Country
Economic Memorandum. World Bank, Washington, DC.
Lachaud, J-P. 2005. “A la recherche de l’insaisissable
dynamique de pauvreté au Burkina Faso. Une nouvelle
évidence empirique.” Document de travail 117, Centre
d’Economie du Développement de l’Université Mon-
tesquieu Bordeaux IV, Bordeaux, France.
———. 2006. “La croissance pro-pauvres au Burkina Faso.
L’éviction partielle de l’axiome d’anonymat en présence
de données transversales.” Document de travail du
département de l’école d’économie du développement,
l’Université Montesquieu Bordeaux IV, Bordeaux, France.
Lipsey, R. G., and K. Lancaster. 1956. “The General Theory of
Second Best.” Review of Economic Studies 24 (1): 11–32.
Massel, B. F. 1970. “Export Instability and Economic Struc-
ture.” American Economic Review 60 (4): 618–30.
McMahon, G., and N. J. Ouédraogo. 2009. “Ouagadougou,
Burkina Faso.” Draft mining chapter for Country Eco-
nomic Memorandum. World Bank, Washington, DC.
CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
Moseley, W. G., and L. Gray, eds. 2008. Hanging by a Thread:
Cotton, Globalization, and Poverty in Africa. Athens, OH:
Ohio University Press.
Newbery, J. 1989. “The Theory of Food Price Stabilization.”
Economic Journal 99: 1065–82.
Pinkney, T. C. 1993. “Is Market Liberalization Compatible
with Food Security? Storage, Trade, Price Policies for Maize
in Southern Africa.” In Agricultural Policy Reforms and
Regional Market Integration in Malawi, Zambia and Zim-
babwe, ed. A. Valdés and K. Muir-Leresche. Washington,
DC: International Food Policy Research Institute.
Platteau, J-P. 1996. “Physical Infrastructure as a Constraint
on Agricultural Growth: The Case of Sub-Saharan
Africa.” Oxford Development Studies 24 (3): 189–219.
———. 2007. “Constraints on African Economic Growth:
The Institutional Legacy.” Paper presented at the first
IERC conference, “The Economic Performance of Civ-
ilizations: Roles of Culture, Religion, and the Law,”
University of Southern California, Los Angeles.
Reardon, T. 1993. “Cereals Demand in the Sahel and Poten-
tial Impacts of Regional Cereals Protection.” World
Development 21 (1): 17–35.
Reij, C., G. Tappan, and M. Smale, 2009. “Re-Greening the
Sahel: Farmer-Led Innovation in Burkina Faso and
Niger.” In Millions Fed: Proven Successes in Agricultural
Development. Washington, DC: International Food Policy
Research Institute.
Rodrik, D. 1996. “Understanding Economic Policy Reform.”
Journal of Economic Literature 34 (1): 9–41.
Savadogo, K. 2009. Le contexte macroéconomique. Back-
ground paper for the Burkina Faso Country Economic
Memorandum. World Bank, Washington, DC.
Stringfellow, R., J. Coulter, A. Hussain, T. Lucey, and
C. McKone. 1997. “Improving the Access of Smallholders
to Agricultural Services in Sub-Saharan Africa.” Small
Enterprise Development 8 (3): 35–41.
Stringfellow, R., J. Coulter, T. Lucey, C. McKone, and
A. Hussain. 1996. The Provision of Agricultural Services
through Self-Help in Sub-Saharan Africa. Research report
R6117CA, Natural Resources Institute, Chatham, United
Kingdom.
SWAC Secretariat, and OECD (Organisation for Economic
Co-operation and Development). 2007. “Economic and
Social Importance of Cotton Production and Trade in
West Africa: Role of Cotton in Regional Development,
Trade and Livelihoods.” Working Paper, Paris.
Teravaninthorn, S., and G. Raballand. 2008. Transport Prices
and Costs in Africa. A Review of the International Corri-
dors. Washington, DC: World Bank.
Tschirley, D., C. Poulton, and P. Labaste, eds. 2009. Organi-
zation and Performance of Cotton Sectors in Africa.
123
 Learning from Reform Experience. Washington, DC:
World Bank.
World Bank. 2004. “Cotton Cultivation in Burkina Faso: A
30-Year Success Story.” Paper presented at the conference
“Scaling up Poverty Reduction, a Global Learning
Process,” Shanghai.
———. 2007a. Burkina Faso: The Challenge of Export Diver-
sification for a Landlocked Country. Diagnostic Trade Inte-
gration Study for the Integrated Framework Program.
Washington, DC.
———. 2007b. World Development Indicators. Washington,
DC: World Bank.
———. 2008. World Development Indicators. Washington,
DC: World Bank.
124 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH
———. 2009a. Burkina Faso: Country Economic Memoran-
dum. Promoting Growth, Competitiveness and Diversifica-
tion. Washington, DC.
———. 2009b. Doing Business 2009. Washington, DC:
World Bank.
———. 2010. World Development Indicators. Washington,
DC: World Bank.
———. Various years. Enterprise Surveys. Washington,
DC.
Yanggen, D., K. Kelly, T. Reardon, and A. Naseem. 1998.
“Incentives for Fertilizer Use in Sub-Saharan Africa: A
Review of Empirical Evidence on Fertilizer Response and
Profitability.” Michigan State University Working Paper
70, East Lansing, MI.
PA R T I I

Post Conflict Situations – Building

Institutions and Governance
 CHAPTER 7
Postconflict Economic Governance Reform:
The Experience of Liberia
Vishal Gujadhur
hen Ellen Johnson Sirleaf began her tenure as
W
president of Liberia in January 2006, she took
charge of a country facing enormous chal-
lenges. It had been more than two years since the signing of
the Accra Comprehensive Peace Agreement that brought an
uneasy end to a conflict that killed more than 250,000
Liberians and left more than 500,000 others displaced. The
scale and intensity of the violence had destabilized the
region politically and economically, making Liberia a global
exemplar of a failed state.
Chief among the president’s many challenges was restor-
ing public trust in economic governance. As the country
collapsed during the conflict, so did government finances
and management of those finances. By 2005 revenue was
less than $95 million a year, with public spending only
about $25 a year on each citizen, one of the lowest levels in
the world. Corruption was rampant. Liberia’s external debt
ballooned to $4.7 billion, or roughly 800 percent of GDP
and 3,000 percent of export in 2005. Domestic debt and
arrears exceeded $600 million by 2006, largely due to weak
government spending controls that allowed ministries and
agencies to spend without central authorization. Economic
mismanagement and the uncertainty surrounding public
finances was reflected in the continuation of UN sanctions
on the export of Liberian diamonds and timber.
The low level of trust in the government’s economic
management capabilities was even more significant given
Vishal Gujadhur is a former Center for Global Development Scott Fellow at the Ministry of Finance in Liberia.
that poor economic governance during the 1970s and
1980s was one of the causes of the conflict. Corruption
and mismanagement of the economy worsened dramati-
cally during the war and prolonged the conflict, as cor-
ruption grew out of control and there was little political
will for a well-run government. Sound economic policies
were therefore, crucial for Liberia’s long-term prospects
for peace.
Initial steps toward reform of economic governance in
Liberia were taken during the transitional period between
the end of the civil war and start of the new administration.
Among these was the launch of the Governance and Eco-
nomic Management Assistance Program (GEMAP) several
months before the January 2006 elections. A partnership
between the government of Liberia and a group of interna-
tional donors, GEMAP was intended to improve the
accountability and transparency of fiscal management in
Liberia. President Sirleaf continued the implementation of
GEMAP after her election, and in addition she committed
to making further economic governance reforms. Five years
on, Liberia’s recovery in terms of public finances and
responsible public policy is striking. Annual government
revenue has nearly tripled since the start of the Sirleaf
administration and substantial inroads have been made in
the budgeting and expenditure processes. Civil servants are
now paid on time, the government has accumulated no new
domestic arrears, and it has reached the completion point in
127
 the enhanced Heavily Indebted Poor Countries (HIPC)
process, allowing more than $4 billion of external debt to be
written off. Liberia’s international reserves also increased,
reaching $108 million in December 2008 and $312 million
in December 2009.1 At the same time, Liberia has climbed
the ranks of Transparency International’s Corruption Per-
ceptions Index, moving from 137 of 158 countries in 2005
to 97 of 180 countries in 2009. The nation has also adopted
the Kimberley Process Certification scheme for sourcing
diamond exports, established new forestry regulations, and
is now fully compliant with the Extractive Industries Trans-
parency Initiative (EITI). Thanks to the improvements in
economic governance, the UN-imposed ban on diamond
and timber exports from Liberia has been removed.
Liberia’s success at gaining control over public finances
in a postconflict environment has prompted a closer look at
the factors underpinning the turnaround. Establishing a
simple causal story is difficult because two major policy
changes happened nearly simultaneously: the start of Presi-
dent Sirleaf ’s administration and the implementation of
GEMAP. The story is further complicated by a general lack
of data resulting from the realities of a postconflict country
and the difficulty of measuring outcomes related to
economic governance.
Many observers have been quick to credit GEMAP for
the economic governance successes that have taken place in
Liberia in recent years. While GEMAP played a useful role,
it also had shortcomings, and an understanding of the con-
text in Liberia and the efforts of the Sirleaf administration is
necessary to establish a more complete story of the recovery
and the reform effort in Liberia.
THE CONTEXT BEHIND ECONOMIC
GOVERNANCE REFORM
Several major events in Liberia’s recent history lie behind
the country’s need for economic governance reform,
namely, the long period of conflict; a difficult economic
recovery following the conflict; and the start of a new, dem-
ocratically elected administration.
Descent into war and emergence from conflict
Liberia’s economic problems began before the first out-
break of conflict in 1989. Even then, the country was
marked by deep social and economic divisions, and Samuel
Doe’s coup d’état in 1980 began a decade of gross misman-
agement until his assassination in 1989. Between 1987 and
1995, Liberia’s gross domestic product (GDP) fell by an
128 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
astonishing 90 percent, and by the time elections took
place in 2005, average income in Liberia was one-fourth of
what it had been in 1987 and one-sixth of what it had been
in 1979.
Finally, after 14 years of instability covering two civil
wars, the Accra Comprehensive Peace Agreement signed in
August 2003 provided for a National Transitional Govern-
ment of Liberia (NTGL) to guide Liberia toward elections
in late 2005. The country faced enormous reconstruction
needs, and international donors stepped in to help set
Liberia on the path to recovery. The transitional govern-
ment took several initial steps toward addressing economic
governance issues. The chairman of the NTGL signed Exec-
utive Order No. 2, which consolidated all government
accounts at the Central Bank of Liberia (CBL). The
government requested audits of practices under the past
regime. The Results-Focused Transitional Framework
(RFTF), meanwhile, set out a comprehensive reconstruction
framework. The transitional government also partnered
with the World Bank, the International Monetary Fund
(IMF), and the European Community (EC) to work on
issues related to governance, economic revitalization, and
financial management.
Persistent fiscal problems and the
genesis of GEMAP
Following a hopeful start, doubts surfaced at the end of
2004 about the NTGL’s commitment to improving eco-
nomic governance and fighting corruption, while Liberia’s
overall economic recovery remained sluggish. A group of
international partners known as the International Contact
Group (ICG) expressed concern that the lack of economic
progress would threaten peace. Audits commissioned by
the EC in early 2005 revealed widespread corruption, not
just during the time of the Taylor regime but also during
the NTGL. At a review of RFTF progress in May 2005, many
international partners focused on the mismanagement of
public resources as a primary reason for the sluggish recov-
ery, kick-starting negotiations with the NTGL for a more
robust international engagement on public financial man-
agement in Liberia.2
The NTGL was reluctant to submit to international
partners’ recommendations. For the NTGL, the two main
points of contention were the submission of GEMAP for
UN Security Council endorsement and the international
recruitment of the chief administrator of the CBL. Only
after repeated threats of halting all international aid did the
NTGL representatives agree to GEMAP. In the end, explicit
linkage to the UN was removed, although the exit from
GEMAP was now linked to Liberia reaching the Enhanced
HIPC Completion Point. This had the benefit of both tying
the completion of GEMAP to a significant incentive (clear-
ing Liberia’s enormous debt burden) and providing
detailed triggers that could be determined at a later date,
because the triggers in a HIPC program generally relate to
public financial management and goals that are in concert
with GEMAP.
Broadly, GEMAP targeted transparency in public finan-
cial management and revenue collection, expenditure con-
trols, and government procurement and concession
practices. Thus conceived, GEMAP ambitiously sought to
accomplish six objectives: securing Liberia’s revenue base,
improving budgeting and expenditure management,
improving procurement practices and granting of conces-
sions, establishing effective processes to control corruption,
supporting key institutions, and capacity building.
A key innovation undertaken as part of GEMAP was the
system of internationally recruited financial controllers
posted alongside Liberians in key agencies. This system was
meant to enable the controllers to help establish transparent
financial procedures, train and build local capacity from
within the agencies, and report on revenue and spending.
The centerpiece of GEMAP design, for which it is known
internationally, is the cosignatory authority these experts
wielded, which ensured that no major financial transactions
could take place without being scrutinized by both a Liber-
ian manager and an international adviser. The basic
GEMAP goal was to “ensure that all revenues due to the
government are collected, and [that] those revenues are
spent according to a budget,” all in an environment of
increasing transparency.3
Although GEMAP was intended to be an interim plan for
the body of international partners involved in postconflict
reconstruction in Liberia, in practice it provided an embed-
ded control system to maintain a transitional economic
governance framework. The intention was to ensure that
government and donor resources were secured and chan-
neled through the budget, but the overall setup was one of
increased oversight. In the end, it would be up to the gov-
ernment to put in place the internal systems and use the
space that this oversight afforded in order to set up long-
term systems.
Elections and the Sirleaf administration
Liberia’s Economic Governance Steering Committee
(EGSC) held its inaugural meeting in October 2005, one
CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
month after the signing of GEMAP. There was little progress
on GEMAP implementation at that point, because Liberia
was focused on the elections. The IMF, which had cut off
relations with Liberia during the war, began discussions
about providing assistance to the country and possibly insti-
tuting a staff-monitored program.
At her inauguration in January 2006, Ellen Johnson
Sirleaf acknowledged the scale of the challenge of rebuilding
the war-torn country. Toward this end, the administration
announced its “150-Day Action Plan” shortly thereafter. The
plan detailed the objectives and deliverables that both it and
the donor community would assume upon taking office.
The 150-Day Action Plan organized the administration’s
work into four pillars: expanding peace and security, revi-
talizing economic activity, strengthening governance and
the rule of law, and rebuilding infrastructure and providing
basic services.
Revitalizing economic activity and strengthening gover-
nance, in particular, are both closely related to the areas cov-
ered by GEMAP. While the capacity for overlap between the
administration’s 150-Day Plan and GEMAP was certainly
possible—as was the possibility, at the time of GEMAP’s
creation, that the new administration would not want to
work within the GEMAP framework—the scenario played
out otherwise. A passage from President Sirleaf ’s inaugura-
tion speech addresses the issue directly:
“If we are to achieve our development and anti-
corruption goals, we must welcome and embrace the
Governance and Economic Management Program
(GEMAP) which the National Transitional Govern-
ment of Liberia, working with our international part-
ners, has formulated to deal with the serious
economic and financial management deficiencies in
our country.
We accept and will enforce the terms of GEMAP,
recognizing the important assistance which it is
expected to provide during the early years of our gov-
ernment. More importantly, we will ensure compe-
tence and integrity in the management of our own
resources and insist on an integrated capacity build-
ing initiative so as to render GEMAP nonapplicable in
a reasonable period of time.” (Sirleaf 2006)
In short, the new administration recognized that with
scarce government resources and limited capacity, the
international community would play a prominent role in
Liberia’s reconstruction and economic governance
reform.
129
 KEY COMPONENTS OF THE APPROACH TO external agencies to place some checks on corruption in the
ECONOMIC GOVERNANCE REFORM absence of domestic impetus. That said, GEMAP also left
ample room for domestic authorities to implement it in
The simultaneous timing of the GEMAP implementation
either more extensive or more limited ways, partly through
and the change in administration in Liberia is an important
the vagueness of such areas as capacity building, but also in
one, because the government of Liberia is both the driver of
the possibility that GEMAP would enable more resources to
interventions and the target of change. Clearly, this duality
be channeled through the budget.
complicates assessments of causality. In addition, GEMAP’s
The Sirleaf administration approached its governing
broad objectives and multiple actors mean that it is “difficult
plans thoughtfully and outlined its key goals and tactics
to trace the boundaries of GEMAP assistance,” as noted in
from the start. These plans can be traced to the 150-Day
a midterm GEMAP evaluation (Morsiani et al. 2008, 13).
Action Plan, a framework similar to Poverty Reduction
Finally, GEMAP and the change in administration were nec-
Strategy Papers prepared by IMF and World Bank member
essarily complementary from the start, with GEMAP provid-
countries. That framework, which was announced at the
ing a set guideline the government could rely on but would
start of the Sirleaf administration, would continue to form
also be constrained by. Indeed, GEMAP was meant to pro-
the overall strategic governance platform as the administra-
vide a transitional economic governance system and enable
tion embarked on an iterative process that moved from the
long-term systems to be put in place. The key question to
150-Action Plan to an Interim Poverty Reduction Strategy
focus on then is the appropriateness of the GEMAP inter-
Paper in 2007 to a full Poverty Reduction Strategy Paper in
ventions and their effect on government policy initiatives.
2008. The administration also used the poverty reduction
While GEMAP was an extensive initiative, its essential
strategy program as an opportunity to set out annual plans
thrust was clear. As stated in an early World Bank/UN joint
for economic governance reforms. In fact, it could be argued
review, “GEMAP is notable for three key features: the scope
that the government’s iterative planning process, including
and intrusiveness of its key features, the consistency of its
efforts such as the IMF’s Staff Monitored Programs and the
features throughout the negotiating process, and a lack of
Poverty Reduction and Growth Facility, are part of the
detail on funding and implementation” (Dwan and Bailey
GEMAP process insofar as they are linked to the Enhanced
2006, 19). The “intrusive” key feature referred to the cosig-
HIPC Completion Point.
natory authority for state-owned enterprises (SOEs), the
From the start, the Sirleaf administration exerted bold
Ministry of Finance, the Bureau of the Budget, and the CBL;
political leadership and made concerted efforts to work with
the position of chief administrator was created in each of
donors collaboratively and ensure coordination. In particu-
these agencies to be the cosignatory with the international
lar, the administration supported a series of locally sourced
financial expert. Additionally, linking the exit from GEMAP
and targeted interventions in key ministries and agencies,
to Liberia achieving HIPC completion shows the imposition
while making use of donor resources and coordination. Also
of forces outside Liberia on the program, despite GEMAP’s
key in the administration’s plans was its effort to make effec-
mentions of the “full respect for the sovereignty of Liberia.”
tive use of donor resources. Given the ravaged state of the
(Government of Liberia 2005).
country at the end of the war, the administration recognized
Other GEMAP features, such as support for Liberia’s
from the start that the reconstruction of Liberia would
General Auditing Commission and the Public Procurement
require significant external assistance. The challenge was to
and Concession Commission, are also important. Those
channel the resources effectively, no easy task given the mul-
features, however, do not differ substantively from interna-
titude of partners, projects, and methods of funding. Flow
tional postconflict reconstruction programs in other coun-
of donor information was poor, and only in the 2009/10
tries, and much of the work process related to them could be
national budget were preliminary estimates of total aid to
completed outside the general framework of GEMAP.
Liberia collected by the government.
Indeed, the other interventions contain much less detail
The Sirleaf administration introduced a new mechanism
about how they are to be implemented, underlining the cen-
to coordinate government and donor activities: the Liberia
trality of the cosignatory aspect of GEMAP.
Reconstruction and Development Committee (LRDC). The
Establishing space for reform was also a key component
LRDC’s Steering Committee was initially kept small, to
of GEMAP. At the time of its formulation, an optimistic
allow decisions to be made more quickly and easily, and
outcome for GEMAP would have been its full utilization by
resembled a subcabinet group chaired by the president.
an incoming administration. At the least, though, it could
Members of the Steering Committee included four
have been a fail-safe within the government and a way for

130 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA

ministers, each with responsibility for one of the recon-
struction pillars, and representation by the four largest
donors (the United States, United Nations, World Bank, and
the European Commission). Representatives of the Eco-
nomic Community of West African States (ECOWAS) and
of the IMF joined meetings, as did other donors and minis-
ters when appropriate. The Steering Committee met every
two weeks at the start of the Sirleaf administration, reflect-
ing the large amount of work to be done, before moving to
monthly meetings.
The LRDC resembled, in some ways, the Economic Gov-
ernance Steering Committee envisioned by GEMAP. By
adopting an administrative structure that mirrored the sub-
stantive pillars of the government’s own comprehensive
action plan, as well as including key donors, a small group
of decision makers was able to make policy decisions and
ensure coordination across efforts. The LRDC also allowed
problems to be addressed as they emerged, permitting
donors to see and participate in the decision-making
processes of the government and understand the realities
the politicians faced. The strength of this arrangement was
ultimately reflected in the decision in 2009 to merge the
EGSC gatherings into the LRDC.
At their core, both GEMAP and the incoming Liberian
administration’s plans attempted to accomplish the same
goals. GEMAP provided an external oversight and control
function, while the government was responsible for carrying
out the actual reforms. After several years, it is now possible
to investigate the underlying efforts and examine their
gains. As discussed earlier, GEMAP’s objectives were broad
and ambitious, which has made it difficult to take stock of
the program’s effectiveness. Further fuzziness around the
program’s edges is the result of programs undertaken by
GEMAP’s international partners on behalf of the govern-
ment of Liberia but not explicitly outlined within GEMAP.
Given the broad scope of public financial management and
establishing fiscal accountability and transparency, the focus
in the future will need to be on two priority areas: securing
Liberia’s revenue base and establishing an effective expendi-
ture process.
Improving revenue collection
Securing Liberia’s revenue base was a primary initial goal of
both the Sirleaf administration and GEMAP. Indeed, first
on the list of GEMAP components is financial management
and accountability. As described in the GEMAP agreement,
“It is critical to protect the revenue streams of key revenue
generating agencies and institutions, as well as to secure the
CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
revenues from customs duties, import levies and taxes”
(Government of Liberia 2005). For the administration,
procedures to improve tax collection were included in the
150-Day Plan under the banner of “economic revitaliza-
tion,” while increased revenues were central to its ability to
begin to fulfill its campaign promises.
The scale of the revenue collection problem was enor-
mous. In fiscal year 2004/05, the government received only
$75 million in taxes, although it had spent more than
$85 million a year between 2000 and 2005. Corruption
affected public spending at all levels of government. Spend-
ing per capita of only about $25 was already among the low-
est in the world, and for the government to begin showing
peace dividends and deliver urgently needed services to its
citizens, revenue had to be raised quickly. The international
community noted in the GEMAP that progress on increasing
revenue collection was not fast enough under the NTGL, say-
ing that “large leakages were taking place due to difficulties
in fiscal administration, absence of verifiable mechanisms
for financial control, and malfeasance” (Government of
Liberia 2005). GEMAP designers also focused on the need to
enforce the NTGL Executive Order mandating that all rev-
enue due to the government be consolidated by the CBL.
While financial management was weak across govern-
ment, GEMAP noted that “SOEs present particularly large
risk given their weak management and operational struc-
tures and systems” (Government of Liberia 2005). This con-
cern was reflected in GEMAP’s primary instrument for
enforcing financial management and accountability. Per-
haps influenced by the audits of the SOEs that had been
conducted by the EC during the NTGL, GEMAP led to the
imposition of management contracts and international
controllers for several SOEs (the National Port Authority,
Roberts International Airport, the Liberia Petroleum Refin-
ery Corporation, the Forestry Development Agency, and the
Bureau of Maritime Affairs) and backed an adviser in the
Bureau of Customs and Excise and a new post of chief
administrator in the CBL.
In hindsight, the focus of GEMAP on placing experts at
SOEs may have been misdirected, because the SOEs were
somewhat peripheral to the central government’s opera-
tions. The Forestry Development Authority, which would
have been the focal point for forestry revenues, was not as
important after the implementation of UN sanctions on
timber exports. That was also true for SOEs as a whole,
because in the absence of profits that can be turned over as
dividends, there was little financial connection between the
central government and SOEs. In fact, the only transfer to
the government from SOEs during the first few years of the
131
 administration was $1.7 million from the LPRC in 2007. As
a GEMAP evaluation report noted, total revenue collection
by the Ministry of Finance in fiscal year 2006/07 was $139
million—approximately four times the combined gross
revenues of the four SOEs with GEMAP controllers (Dod
and Nelson 2008).
As for the Bureau of Customs and Excise at the Ministry
of Finance, which collects over half of government revenue,
the improvement in revenue performance has not been a
result of GEMAP. An evaluation of the first short-term
expert working in those two agencies found him to be of
substandard quality, and three years after the signing of
GEMAP there had been no systematic donor support under
GEMAP auspices (Morsiani et al. 2008). In fact, government
action and IMF support were much more salient in improv-
ing revenue collection, as evaluations have concluded. A
U.S. Agency for International Development (USAID)
report, however, has noted that while there have been efforts
under GEMAP to support systemic improvements in rev-
enue collection in Liberia, particularly in the customs arena,
the assistance was “not being fully utilized for the benefit of
Liberia” and has produced little follow-up for functions that
had been installed (Dod and Nelson 2008).
How, then, was the Liberian government able to have a
nearly instantaneous effect on revenue collection? The
150-Day Plan announced a target of increasing revenues
by at least 15 percent year-on-year, and indeed, the overall
revenue from international trade increased by 35 percent
in the first five months of the Sirleaf administration (Dod
and Nelson 2008). This was accomplished largely by better
use of solutions already proposed. As the new government
came to office, customs payments were collected in an
“unwieldy and complex manual system,” according to a
GEMAP assessment (Dod and Nelson 2008). To address
this, a technical expert from the U.S. Treasury designed an
electronic computerized receipt system during 2005 in
collaboration with USAID, the Ministry of Finance, and
the CBL. However, the system was not fully used and rec-
onciliation of the CBL/Ministry of Finance account did
not begin until the new administration took office. By the
end of January 2006, before any GEMAP advisers had been
deployed across government, the newly installed Liberian
authorities accepted only computer-produced receipts at
the CBL. Liberian authorities also began to ensure more
stringent preshipment inspections for goods moving
through the port, and began to eliminate noncash tax
payments—previously, the government had accepted “in
kind” payments. This sort of insistence on adherence to
existing rules and use of systems already in place allowed
the government to produce the dramatic 35 percent
increase in revenue collection, and in the process demon-
strated newfound political will.
The government’s success in increasing revenues was not
limited to collection of customs duties; in fact, it was real
and sustained across several areas of the government. Cus-
toms duties nearly doubled in the first year of the Sirleaf
administration, from $35.3 million in fiscal year 2005/06 to
$69.9 million in 2006/07. This is all the more impressive
considering that the percentage increase in imports during
this time was much smaller (imports rose from $294 million
in 2005 to $418 million in 2007). The government also
began to reorganize its methods for collecting income taxes.
Efforts in this area paid off immediately, with actual tax rev-
enues increasing 19 percent between February and June
2006 (IMF 2006). GEMAP experts did not arrive in Liberia
until mid-April 2006.
Table 7.1 shows that the increase in Liberian revenues was
not limited to customs duties (shown in the table as interna-
tional trade taxes). The deputy minister of revenue of the
Ministry of Finance, in tandem with the IMF, introduced a

Table 7.1 Revenue Collection in Liberia

(millions of dollars)
Revenue 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09
Total revenue 56.0 79.3 84.6 146.8 200.8 211.3
Total tax revenue 55.3 75.7 81.0 140.0 168.8 190
International trade tax 24.3 30.2 35.3 69.9 79.1 87.9
Income tax 9.4 28.4 25.1 42.5 52.6 65.8
Goods and services tax 21.3 16.9 20.3 26.1 34.9 33.7
Other 0.3 0.3 0.3 1.4 2.2 2.6
Nontax revenue 0.8 3.6 3.6 6.9 32.0 21.4
From SOEs 0 0 0 1.7 0 0
Grants 3.0 1.0 1.0 1.5 0 23.6

Source: IMF Liberia Program Reviews.

132 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA

series of actions aimed at both short- and medium-term
revenue collection. A partial list of actions includes com-
puterization of the tax payment system; strengthening cus-
toms administration and steadily reducing exemptions;
increasing penalties for not undergoing preshipment
inspection, reorganizing the tax collection system into large,
medium, and small taxpayer units, with an emphasis on the
large taxpayer units; and increasing the excise tax on beer
and cigarettes. The turn away from exemptions and short-
term fixes that had been common during the periods of
instability in favor of modernization of tax collection and
rules-based collection is reflected in an increase in the
income tax revenue take of more than 50 percent during the
first year of the Sirleaf administration.
Another central platform in the GEMAP effort to
increase revenue collection in Liberia was the institution of
the chief administrator position at the Central Bank of
Liberia. The original idea was for the administrator to serve
under the guidance of the governor of the CBL, to possess
binding cosignatory authority on operational and financial
matters, to advise on banking operations and to emphasize
internal controls and audits. These plans never came to
fruition, however, because the governor of the CBL consid-
ered the cosignatory arrangement unacceptable and the
first GEMAP adviser found it difficult to work within the
system and left the position after 15 months. After this, a
GEMAP adviser was placed at the CBL in a revamped posi-
tion, focused more on operational matters.
The second GEMAP adviser was able to contribute sig-
nificantly to improvements in the CBL system. The CBL was
able to register surplus in its financial position, introduce a
travel policy harmonized with the rest of the government,
and increase its dollar reserves from $6.4 million at the start
of 2006 to $35.1 million in 2007. The midterm GEMAP
evaluation, though, recommended “a careful scrutiny in the
effectiveness of GEMAP assistance to CBL given the fact
that the conflicts it creates are not conducive to the achieve-
ment of results” (Dod and Nelson 2008).
Despite their successes, GEMAP efforts were a relatively
small part of the overall task of improving revenue collec-
tion in Liberia, primarily because of GEMAP’s targeting of
SOEs. During the war and the NTGL period, resources con-
tinued to flow to SOEs and agencies, while the central gov-
ernment remained weak. Centralization of government
accounts significantly reduced opportunities for leakage.
Although GEMAP’s support to SOEs was useful in improv-
ing those institutions’ performance, it was peripheral to the
revenue recovery led by government efforts and non-
GEMAP-related donor support. This finding is echoed in
CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
evaluation reports, which found that GEMAP performed
“unremarkably” in regard to the improvement of central
banking and securing revenue at the Bureau of Customs
(Morsiani et al. 2008).
Establishing an effective expenditure process
Just as important as improving revenue collection in Liberia
was increasing the transparency and effectiveness of public
spending. One of the stated aims of the government was to
“restore the government’s credibility in using the country’s
scarce resources efficiently and effectively,” while a primary
GEMAP objective was to improve the controls in public
expenditures. Gains in effective expenditure, however, are
more difficult to isolate than gains in revenue collection,
because metrics of the quality and efficiency of spending are
generally not objective or are unavailable in Liberia.
During the NTGL period, the meager amounts of rev-
enue the government collected were not spent effectively,
and reports were widespread in the Liberian media of cor-
ruption and similar accusations among various factions in
the NTGL. Because the Bureau of Budget was initially oper-
ationally independent from the Ministry of Finance, the
structure required cooperation and collaboration in budget
formulation and execution. Past governments, meanwhile,
had kept the national budget secret and allowed unlimited
transfers between line items at the executive branch’s whim.
Procurement did not follow standardized procedures. In
short, the structure the Sirleaf administration inherited did
not fit its aims.
As an interim measure, foreign policy advice and techni-
cal assistance during the NTGL established a cash-based
balanced budget with budget execution handled through
the Cash Management Committee (CMCo), which consisted
of the Bureau of Budget and Ministry of Finance officials,
who approved all government expenditures. To introduce
better procurement practices, international partners helped
draft a new Public Procurement and Concessions Act. By
the time of the 2005 IMF Article IV consultation, however,
IMF staff found that “significant expenditures were still
made without the committee’s prior authorization” and that
the CMCo was “not operating as intended” (IMF 2005).
Reviews of the Ministry of Finance found that it “operated
in constant crisis mode, with the CMCo under constant and
enormous pressure to clear payments,” while “poor docu-
mentation, thirty stages of voucher review, lack of informa-
tion and opaqueness regarding prioritization exacerbated
[problems]” (Morsiani et al. 2008). With no reliable esti-
mates of budgetary receipts and balances, it was hard for
133
 line ministries to learn of their budgetary allotments, and
the lack of commitment controls led to a buildup of gov-
ernment liabilities. By the end of 2004 the government had
approximately $12 million in domestic arrears, roughly
15 percent of the total budget.
GEMAP interventions aimed at improving budgeting
and expenditure management consisted of the establish-
ment of an Integrated Financial Management Information
System and the placement of advisers with cosignatory
authority in the Bureau of Budget and Ministry of Finance.
The incoming government also took action in this area,
indicating in its 150-Day Action Plan that the CMCo should
be effectively operational upon the start of the new admin-
istration in January 2006. In this case, the work done by
GEMAP and the government were mutually reinforcing.
Expenditure controls first took root through the CMCo,
which quickly began to function effectively and arguably
became one of the successes of GEMAP and the govern-
ment. Under GEMAP’s commitment control system, all
purchase orders and expenditure vouchers from line min-
istries were submitted to the CMCo, which verified that
each proposed expenditure was consistent with the spend-
ing agency’s monthly expenditure allocation and compliant
with the procurement law. The CMCo also ensured that
there were sufficient funds in the government’s accounts at
the CBL. The system in which payments required CMCo
authorization prevented expenditure arrears and unautho-
rized government purchases. Although similar systems had
been in place beforehand, they were cumbersome, had little
political support, and consequently were rarely enforced.
In a September 2006 review, IMF staff found that
Liberia’s commitment control system had been well imple-
mented. From the start of the administration, no new
arrears accumulated and all expenditure authority was con-
solidated with the CMCo. Through public notices, the gov-
ernment made it clear that commitments not accompanied
by CMCo vouchers were not legal claims on the govern-
ment. Although the foreign adviser at CMCo likely played a
significant role, CMCo also had a foreign adviser during the
NTGL period. What changed with the Sirleaf administra-
tion was not the presence of a foreign adviser but the polit-
ical environment and support for that adviser.
Within the Bureau of the Budget, complementary
improvements were implemented. Even before the
GEMAP-recruited adviser began to work with the Bureau
of the Budget in April 2006, there was increased communi-
cation with ministries to ensure that commitments did not
exceed available revenues. The task of the Bureau of the
Budget, however, was complicated by the importance of
134 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
both short- and medium-term targets: the monthly allot-
ments and current budget had to be executed appropri-
ately, while the following year’s budget had to be prepared
simultaneously. The arrival of the GEMAP adviser brought
increased focus on both immediate controls in terms of
cosigning of allotment requests and capacity building
through training of staff at the Bureau of the Budget. The
importance of the adviser was also discernible in reports
noting instances of abuse when the adviser was temporar-
ily absent (Dod and Nelson 2008).
In the lead-up to the merger of the Bureau of the Budget
and the Ministry of Finance, reactivation of Liberia’s inter-
agency National Budget Committee led to improved coordi-
nation and oversight of both budget policy formulation and
expenditure. In addition, the budget preparation process
became more open and participatory, particularly during
the fiscal year 2007/08 budget cycle, when both the GEMAP
adviser and local staff began to focus on long-term planning
and processes. This sequenced intervention—first providing
for the immediate stabilization of the procurement system
and then moving toward long-term planning—was an
example of effective implementation of the GEMAP.
A key enabling factor for Liberia’s success in establishing
an effective expenditure program was senior management
at both the Ministry of Finance and Bureau of the Budget.
Given the political nature of a budget in a democracy, this
also involved strong links with Liberian civil society, and
the lengths to which the administration went to work with
civil society were rewarded with a relatively high Public
Expenditure and Financial Accountability (PEFA) score of
“B” by 2007.4
While Liberia should be lauded for its success in manag-
ing its budget and expenditures, it should also be noted that
the Integrated Financial Management Information System
did little to improve Liberia’s budget and expenditure
reform process. As of 2010 the system still had not been
implemented. On the other hand, the Resource Manage-
ment Unit, the pool of donor money originally intended to
fund technical assistance and support for the system, was
able to be used to attract talent to and finance a manage-
ment training school and an aid management unit within
the Ministry of Finance. In fact, the RMU became an
adaptable method of ensuring that the Ministry of Finance
received technical assistance for its changing needs and
could fill immediate capacity gaps where advisers had
direct operational responsibilities. With this flexible setup,
Liberian leaders were able to implement their own pro-
grams and initiate the day-to-day operational work needed
to get Liberian institutions running effectively again.
 Overall, evaluation reports are favorable for GEMAP on
securing and stabilizing revenue collection and improving
expenditure management, despite the fact that the processes
were not fully in line with what was intended at program
conception. That said, the GEMAP-appointed advisers at
the Ministry of Finance and Bureau of the Budget had a
positive impact, and reports suggest cosignatory authority
was useful in these cases, although the advisers were also
significantly enabled by the leadership in various min-
istries. In both expenditure management and budget coor-
dination, GEMAP was able to play a role in stabilizing the
system and creating capacity for more significant reforms to
take place. How this capacity would be used, however, is
another question. Several additional accomplishments
made under GEMAP are illustrated in box 7.1.
THE LINK BETWEEN THE GOVERNMENT
OF LIBERIA AND GEMAP
While it is difficult to assess how the Liberian government
would have performed in the absence of GEMAP, it is clear
that without government buy-in, GEMAP would have oper-
ated very differently. In the case of revenue collection, the
Liberian government acted almost independently of
GEMAP, driving the reform process, whereas in establishing
expenditure and budgeting controls GEMAP had a larger
role in helping the government stabilize the system. The
commitment and actions of the government in Liberia to
the changes in public accountability are striking, yet many
stock-taking reviews have focused only on the role of
GEMAP. Any overall report card for the administration’s
performance, on economic governance or otherwise, is
Box 7.1 Other Accomplishments under GEMAP
GEMAP consisted of six pillars. In addition to the pillars
focused on improving revenue collection and establish-
ing effective expenditure processes, the GEMAP pillars
focused on supporting key institutions and establishing
effective processes to control corruption.
These two pillars had contrasting outcomes. In the
first case, the evaluation teams appointed by GEMAP
to support the General Auditing Commission were
successful, playing a strong role in rebuilding institu-
tional and capacity building. The European Commu-
nity provided long-term technical support, financed
an auditor general position (currently held by a
CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
perhaps complicated by the government of Liberia being
both an agent and target of change. Still, the effect of the
change in political leadership in the sequence of reform is
undeniable. Figure 7.1 provides an overview of the timeline
of economic governance reforms in Liberia.
Without a doubt, the shift between the NTGL period
and the Sirleaf administration in terms of economic man-
agement was marked. President Sirleaf was democratically
elected, selected members of her government, and had a
comprehensive plan to rebuild Liberia. She signaled a clear
break with the policies of the past, and one of the first
agenda items in the new administration’s 150-Day Action
Plan was to dismiss all NTGL political appointees from the
Ministry of Finance. She also endorsed GEMAP in her
inaugural speech and included GEMAP implementation
benchmarks in the 150-Day Action Plan. Political leader-
ship, then, was clearly important to both the functioning of
GEMAP and the change in economic governance. GEMAP
evaluators recognized this. Indeed, without this change in
leadership, GEMAP’s operations would necessarily have
been vastly different.
All of this said, the openness of Liberian politicians to
receiving technical assistance and accepting foreign advisers
working alongside them is also striking. In part, it reflects
the scale of work that needed to be accomplished. It likely
also reflects a country in which both the president and the
minister of finance were international technocrats with sig-
nificant policy-making experience. Either way, it made for a
collaborative process that helped to drive economic gover-
nance reform forward.
Government buy-in was particularly important in
Liberia’s case because at its core, GEMAP established only a
Liberian) and the refurbishment of the General
Auditing Commission, and provided general advisers.
The result has been an independent General Auditing
Commission that provides, in a very public manner,
an ex post check on fraud and support for accounta-
bility.
By contrast, the Liberia Anti-Corruption Commis-
sion has yet to take off. Although donors explicitly out-
lined an Anti-Corruption Commission in the GEMAP
document, the legislation to establish such a commis-
sion did not pass until 2008, and then donor funding
support proved to be lacking.
135
 Figure 7.1 Economic Governance Reform Timeline

250
August 2003: Civil September 2005: NTGL November 2005:
war ended by reluctantly signs GEMAP April 2006:
Ellen Johnson
Accra after international First GEMAP
Sirleaf wins run-off
comprehensive concern over corruption controllers arrive
200 election
peace agreement leads to pressure

150

100

January 2009: First

EITI report
50 May 2007: Budget published; Liberia
January 2006: June 2006: submitted to legislature March 2008: soon becomes first
President Sirleaf Customs receipts as per agreed time, Liberia reaches African country
assumes office, up by 35% yoy in followed by HIPC decision designated
initiates 150-day plan first 150 days participatory hearings point EITI-compliant
0
2003 2004 2005 2006 2007 2008 2009

GoL revenue collection

Note: Revenue data is exclusive of grants and is from the IMF. Liberia’s fiscal year runs from July-to-June, all figures reported are for the fiscal year ending
in June (i.e., 2003 is for the fiscal year 2002/2003)

transitional economic governance framework that merely
created the space for the government to carry out the more
operational aspects of reform. While a GEMAP-style pro-
gram might work in the case of a hostile recipient govern-
ment, such a situation is unlikely. One program, the
USAID-sponsored “Liberia Economic Stabilization Support
Program,” which strongly resembled GEMAP, can serve to
inform the feasibility of GEMAP under a different gover-
nance regime. Implemented in 1989, that program sought
to improve revenue collection and expenditure control by
giving 17 experts financial control over government
accounts (Dwan and Bailey 2006). Its failure in less than a
year underscores the importance of dynamic leadership and
political will.
GEMAP certainly complemented the government’s
efforts and put in place a stop-gap system, but questions still
surround the mix of its activities. Although GEMAP’s inten-
tion, in part, was to lay the foundations for reform, the
activities carried out under it concentrated on keeping the
system moving forward as constructed. The implementa-
tion of the CMCo and financial controllers, for example,
while successful, left most of the heavy lifting in terms of
long-term building of systems to the government. So, while
the interim controls carried out under GEMAP were suc-
cessful in helping to prevent new arrears and rooting out
opportunistic corruption, their effectiveness resulted in a
top-heavy structure that led to a difficulty in allowing a
more successful rollout of spending.
By 2008 it was apparent to both the government and its
international partners that a lack of capacity at the line min-
istry level was hindering the effective implementation of the
government’s agenda. Difficulties implementing the pro-
curement law led to a slowdown in critical government
spending that GEMAP’s stop-gap, concentrated measures
were not equipped to address. The capacity building that
GEMAP was supposed to deliver could possibly have
removed this bottleneck but did not materialize in a signif-
icant way. Instead, the focus on the immediate controls
reduced the long-term relevance and effectiveness of
GEMAP, leaving the government to drive the capacity-
building agenda with donors.
In addition to enabling GEMAP to be a success, the polit-
ical leadership in Liberia also took independent actions to
improve financial management and transparency from the

136 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA

start. In her inaugural speech, President Sirleaf promised
that there would be actions taken “so as to render GEMAP
unnecessary,” implying that she intended for the new
administration to target not just GEMAP actions but also
the larger economic governance areas that GEMAP covered.
In the end, the Liberian government’s drive for reform out-
grew the outlines of GEMAP assistance once the system
had stabilized.
FURTHER CAPACITY FOR ECONOMIC
GOVERNANCE REFORM
Despite significant progress since 2005, Liberia is still a frag-
ile state, and much work remains to be done. That said, crit-
icism of the government or of donor interventions should
not downplay the enormity of the tasks before it just five
years ago, and many of the concerns today were not of top
priority during the initial launch; when stabilizing the econ-
omy was the government’s unique focus.
Liberia’s most problematic economic governance short-
fall continues to be capacity building, the most vaguely
defined objective in the GEMAP program document, which
noted the other components would be accompanied by “a
plan and resources to enable major progress for capacity
building” (Government of Liberia 2005). In general, capac-
ity building has never been a priority under the GEMAP
program. A two-year capacity-building training program at
the Ministry of Finance, though useful, is the only explicit
capacity-building initiative conducted under GEMAP
auspices. More informal programs took place in other
ministries.
One GEMAP evaluation reported that even though line
ministries and agencies were committed to economic gover-
nance improvements, their lack of capacity severely limited
their performances. This shortcoming was exacerbated by a
new public procurement law that, while necessary, decentral-
ized procurement without first building capacity to facilitate
its implementation. Overall, however, the will to retrain
Liberian government workers on new procedures and regu-
lations has been lacking, and the role of GEMAP in capacity-
building initiatives has not always been constructive.
In hindsight, GEMAP should have paid closer attention
to capacity building from the start. Rather than allowing its
focus to pivot as needed based on the situation in Liberia, it
appears that GEMAP efforts focused on form, rather than
function, with too much focus on SOEs, despite GEMAP
reviews suggesting a change in focus. One evaluation noted
that, “When best practices are in place, cosignature arrange-
ments are not a function of control but complement capac-
CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
ity and institutional building” (Morsiani et al. 2008) and
recommended a transition from a focus on immediate con-
trols to longer-term system building.
The Sirleaf administration made its views clear early on.
In a paper prepared for the Liberian Partners Forum in Feb-
ruary 2007, the administration gently noted that donors
continued to focus on the short term: “Previous break-
downs in governance led to considerable focus being placed
on the ‘cosignatory’ aspects of GEMAP. Considering the
progress and positive changes that have taken place in the
leadership of the country, [the government of Liberia] has
encouraged partners to focus on the development of atti-
tudes, systems and procedures that can promote and sustain
integrity, rather than on a transient policing role” (Govern-
ment of Liberia 2007).
While this focus on stabilization was appropriate and
useful to economic governance reform in an immediate
postconflict situation, a second phase was needed to deepen
upstream and downstream controls on spending, including
building up internal and external audits and increasing the
public procurement capacity in spending agencies, a task
that GEMAP frequently bypassed. The controls introduced
under GEMAP were at best transitory, and it is possible that
they were actually a hindrance to establishing deeper, more
strategic reforms. Indeed, despite GEMAP efforts to
strengthen systems, most donors have yet to trust Liberia
with budget support. Like GEMAP’s transitional systems,
donors continue to intervene in Liberia in an ad hoc fash-
ion, from providing project support to multidonor trust
funds to contributing to a variety of parallel implementa-
tion units designed to work around the Liberian system
rather than build it up.
As further confirmation of the need for sustained
capacity-building efforts in Liberia, the national capacity
development plan currently being formulated under the
auspices of GEMAP (the same report discusses progress on
the implementation of Liberia’s poverty reduction strategy)
finds that capacity building is the primary constraint to
faster progress on economic governance reform in Liberia.
Without the capacity to implement, the long-term systems
upon which sustained success depends will remain elusive.
KEY LESSONS LEARNED FROM THE LIBERIAN
ECONOMIC GOVERNANCE EXPERIENCE
Despite the difficulties in attributing causality to gover-
nance policy interventions given the two policy changes that
took place in early 2006, it is possible to draw lessons from
Liberia’s public financial management gains.
137
 ■ Placing international advisers in postconflict environ-
ments can be useful in some cases. The GEMAP pro-
gram was primarily aimed at top-down controls, and this
target was appropriate and valuable in the Ministry of
Finance and SOEs in the initial transitional NTGL period
given the situation at the time. Building up systems, how-
ever, requires more than top-down checks on abuse. For
the most part, stationing advisers in the ministries both
helped Liberia fill immediate capacity gaps and estab-
lished controls and set examples for those willing to fol-
low them. This was dependent, however, on local author-
ities being open to the process. The cosignatory process
was useful in some cases, but GEMAP tended to be a
focus on the process itself, rather than the function the
reforms were to intended to carry out.
■ Finding the appropriate role for international partners
is key. In 2005 a key question in the process of economic
governance reform appeared to be whether discipline
could be imposed from the outside. While it is simplis-
tic to view governance interventions on a continuum
between international trusteeship and full country own-
ership, most people familiar with the situation in Liberia
considered that a significant amount of external assis-
tance was needed. In the case of Liberia, however, the
government had changed by the time GEMAP was for-
mulated, complicating assessment of how effectively
GEMAP fulfilled its intentions. Indeed, GEMAP helped
improve economic governance in Liberia, but only after
the new administration changed the parameters of its
operation.
■ Partners must strike a balance between capacity build-
ing and delivery. Under GEMAP, international con-
trollers were to serve two roles in Liberia. First, they were
to impose temporary controls and give the government
the necessary breathing space to put long-term systems
into place. Second, they were to help fill in for the lack of
capacity and help to build up the long-term systems
from the inside. While those two roles are interrelated,
the distinction between the two continues to matter.
GEMAP was ostensibly designed to cover both, but the
focus ended up being on delivery. Once Liberia’s top-
level governance problems were resolved, capacity
became the binding constraint to fully rebuilding gov-
ernment institutions, and by some accounts, GEMAP did
not do enough to address that.
■ Provisions must be made for flexibility. The timing of
GEMAP, which was signed during the transitional gov-
ernment in Liberia, required that donors be flexible in
their support of the incoming government. An economic
138 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
governance reform program should, for example, be able
to provide for rapid-progress scenarios. Although
GEMAP met with success in several of its interventions,
it focused on the control process and did not go much
beyond it, which lessened its impact in terms of building
institutions that go beyond the initial stabilization of
the system.
In terms of revenue collection, for example, the initial
GEMAP interventions targeted SOEs, which were impor-
tant but not especially relevant for long-term fiscal gains in
Liberia. More generally, as pointed out in one evaluation,
“assistance provided by GEMAP partners outside GEMAP
arrangement were . . . more relevant to the specific needs
of the various institutions” (Morsiani et al. 2008).
■ Supporting local leaders can improve economic gover-
nance outcomes. Ultimately, success in economic gover-
nance reform depends on local efforts, and leadership in
particular was a key ingredient in Liberia’s success. Addi-
tionally, local leaders took charge of programs through
constant dialogue with donors and took ownership of
the reform process. Although it was impossible to know
at the time GEMAP was signed whether the new admin-
istration would fully support the program, the new
administration found GEMAP to be, in the words of
President Sirleaf, “a necessary intrusion.” Indeed, the
reform progress under GEMAP was only a start for the
government, and the impetus provided by the new gov-
ernment in addressing economic governance, imple-
menting strong expenditure controls, and securing and
expanding revenue cannot be underestimated. In the
end, strong leadership may be both the most relevant
point and the hardest to replicate.
NOTES
1. Special SDR allocation included.
2. For more detailed discussions on the genesis of GEMAP,
see Dod and Nelson (2008).
3. Retrieved from http://www.gemapliberia.orsg/pages/
accomplishments.
4. www.pefa.org.
REFERENCES
Dod, D., and E. Nelson. 2008. “USAID Activities under
GEMAP in Liberia: Impact Assessment Report.” U.S.
Agency for International Development, Washington, DC.
Dwan, R., and L. Bailey. 2006. “Liberia’s Governance and
Economic Management Assistance Program (GEMAP): A
 Joint Review by the Department of Peacekeeping Opera-
tions’ Peacekeeping Best Practices Group and the World
Bank’s Fragile State’s Group.” United Nations, Department
of Peacekeeping Operations, New York.
Government of Liberia. 2005. “Government and Economic
Management Assistance Program.” Monrovia.
———. 2007. “Progress under GEMAP.” Monrovia.
IMF (International Monetary Fund). 2005. “Liberia: Article
IV Consultation: Staff Report.” Country Report 05/166.
Washington, DC.
CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA
———. 2006. “First Review of Performance under the Staff-
Monitored Program.” Country Report 06/412. Washington,
DC.
Morsiani, G., et al. 2008. “Mid-term Evaluation of the
Governance and Economic Management Assistance
Program.” GEMAP, Monrovia.
Sirleaf, Ellen Johnson. 2006. Inauguration Speech Transcript.
www.emansion.gov.lr/doc/inaugural_add_1.pdf).
139
 CHAPTER 8
Decentralization in Postconflict Sierra
Leone: The Genie Is Out of the Bottle
Vivek Srivastava and Marco Larizza
hen Sierra Leone emerged from more than a
W
decade of conflict in 2002, it was one of the
poorest countries in the world. It faced huge
development challenges, with much of its infrastructure
having been destroyed during the war.
The areas outside Freetown had traditionally been
excluded and marginalized. In fact, the overcentralized sys-
tem of rule, which excluded the majority of the population,
was one of the key causes of the conflict. There were internal
and external pressures on the government to be more inclu-
sive to establish its legitimacy and reverse the conditions
that led the country to conflict (Truth and Reconciliation
Commission 2004; Hanlon 2005; Kieh 2005).
The reestablishment of local governments, through the
Local Government Act of 2004, was an important initiative
in this direction undertaken by the Sierra Leone People’s
Party (SLPP) government of Abdul Tejan Kabbah. The leg-
islative framework provided by the act and the associated
regulations for political, fiscal, and administrative decen-
tralization (with some exceptions) provided a robust foun-
dation for the establishment of decentralization through
devolution of key functions from the central government to
local councils. They also provided a simple and easily
understood system for intergovernmental transfers.
With financial support from donors and through the
efforts of a set of donor-supported agencies established
within government, a functioning system is now in place.
Although the process has been somewhat slower than
desired, there has been a steady devolution of functions and
finances, and local councils are now fully staffed. Each coun-
cil has a core staff of development planners, internal audi-
tors, monitoring and evaluation officers, and procurement
officers with requisite capacity for managing their service
delivery functions. Both the legislation and its implementa-
tion leave open the possibility, however, of the center dom-
inating and manipulating the subnational governments by
playing off the traditional authorities (chieftaincies) against
the local authorities through a “divide and rule” strategy
and by minimizing the autonomy of the local councils over
the control of financial and human resources.
This chapter examines the devolution of power in Sierra
Leone since the end of armed conflict. The first section
briefly reviews the history of decentralization in Sierra
Leone and discusses the incentives and motivations that
may have influenced the government’s decision to decen-
tralize in 2004. The second section highlights the key fea-
tures of fiscal, administrative, and political decentraliza-
tion by comparing the legal (de jure) provisions of the Local
Government Act with the actual (de facto) implementation
experience during the period 2004–10. The third section
summarizes the major achievements of decentralization to
date, focusing on the impact on service delivery and local
governance. Addressing the politics of decentralization, the
fourth section identifies potential threats and emerging evi-
dence that suggests that the national government may be
trying to regain control and manipulate local politics in a
way that would be optimal for the center. The last section
summarizes the main arguments and suggests politically
141
 feasible options available for development partners to move
the decentralization agenda forward and prevent a reversal.
THE LEAD UP TO THE ESTABLISHMENT
OF LOCAL GOVERNMENTS
Under colonial rule, the British authorities established a
strong political and administrative divide between the
Colony (Freetown and the Western Areas) and the Protec-
torate (the rest of the country, which was divided into three
provinces) (figure 8.1). The Colony was ruled by an elected
local government and a British governor representing the
monarch. The Protectorate was administered through a
system of “indirect rule” in which traditional authorities
(the chieftaincy) were appointed by the state for the collec-
tion of revenue, the maintenance of law and order, and the
resolution of local disputes.
After independence, outside of Freetown the local
administration retained many of the features of the British
colonial indirect rule system, but elected local councils also
functioned until 1972, when President Siaka Stevens abol-
ished them, moving their responsibilities to the central gov-
ernment (Fanthorpe et al. 2006). Management committees
superseded local town councils, but they focused largely on
the collection of market dues. They became grossly over-
staffed, with employment provided as a reward for political
support, and accountability for service delivery was under-
mined. This dramatic retraction of local government, which
coincided with the period of the All People’s Congress
(APC) single-party rule, meant that traditional authorities
represented the only form of governance in the provinces.
Although chiefs became increasingly active in the collection
of revenues on behalf of the central state, they had no ser-
vice delivery or development functions.
This reliance on chiefs, who were often repressive, and
a lack of formal state structures in most of Sierra Leone
led governments to start considering decentralization as a
viable option to mitigate popular discontent. Following
multiparty elections in 1996, the new government prepared
a national document entitled the “Good Governance and
Public Sector Reform Strategy.” The document focused on
decentralization as a major instrument for reform, espe-
cially in deprived rural areas. It suggested that decentral-
ization could help rebuild service delivery and improve
quality; improve resource allocation, by moving resources
to the service delivery level; ensure greater citizen engage-
ment in the processes of government; and involve the
community in its own development by enhancing citizens’
participation in the planning and implementation of
development strategies and the setting of priorities.

Figure 8.1 Territorial Organization of Sierra Leone

a . Pr o v in c e s an d t h e W e s ter n A r e a b . D i s tri cts a nd chiefdoms

IBRD 38553
MAY 2011
10° 13° 12° 11° 10° 10° 13° 12° 11° 10°
GU INE A GUINEA GUINEA GUINEA
Kabala Kabala
KOINADUGU
BIA BOMBALI
Kambia NORTHERN K AM
Kambia
9° 9° 9° 9°
Makeni Makeni
Port Loko Port Loko KONO
Magburaka Magburaka
Sefadu PORT
FREETOWN FREETOWN
LOKO TONKOLILI Sefadu
URBAN
WESTERN RURAL
Kailahun WESTERN Kailahun
Moyamba
EASTERN AREA Moyamba
BO UN
LAH
MOYAMBA KAI
8° Bo 8° Kenema 8°
Bo
SOUTHERN Kenema
KENEMA
Bonthe
ATLANTIC
Bonthe
ATLANTIC BONTHE Pujehun
Pujehun LIBERIA PUJEHUN
LIBERIA
OCEAN OCEAN
7° 13° 12° 11° 7° 7° 13° 12° 11° 7°

Source: Zhou 2009.

Note: The Republic of Sierra Leone is composed of three provinces (the Northern, Southern, and Eastern Provinces) and a region known as the Western
Area, which is governed by a rural council and a city council for Freetown, the nation’s capital. The provinces are divided into 14 districts, which are divided
into chiefdoms.

142 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
 In 2004 the government embarked on a nationwide DECENTRALIZATION IN PRINCIPLE
decentralization reform program, with the official goal of AND IN PRACTICE
addressing some of the root causes of civil war and
The Local Government Act and its statutory instruments
improving the delivery of basic services. The decentraliza-
provided the legislative framework governing decentraliza-
tion program was officially launched in 2004, when the
tion in Sierra Leone. With some exceptions, the act laid
Parliament passed the Local Government Act and its
down a robust framework for political, fiscal, and adminis-
related statutory instruments. The process of decentraliza-
trative decentralization in Sierra Leone. The framers of the
tion was designed around two major components: the re-
document took a pragmatic approach, putting in place a
creation of the local councils, which had been abolished in
framework that included “transitional” provisions that were
1972; and an attempt to re-create and re-legitimize the
acceptable in the absence of a policy on decentralization.
institutions of the chieftaincy, which had suffered greatly
Policies on decentralization and chiefdom governance were
during the period of one-party rule and the civil war. The
to be prepared subsequently. The act recognized the need
government and its development partners viewed the
for change and identified the period up to 2008 as the first
revival of subnational political institutions—which had
phase of the arrangements for fiscal decentralization.
existed before the period of one-party rule—as a primary
An important area on which the act did not bring closure
strategy with which to build popular legitimacy, sustain
is the relationship between the local councils and the chief-
political stability, and reverse the massive urban/central
taincy. Although by law the chiefdoms are subordinate to
bias that was a feature of Sierra Leone’s postcolonial poli-
the local councils, chiefs have not accepted this hierarchy,
tics and led the country into war.1
and ambiguity on the part of the national government per-
Although the efficiency gains normally associated with
sists. In particular, the issue of the revenue domain was not
decentralization may have played a role in determining
satisfactorily resolved with respect to the local tax. Under
the government’s decision to decentralize, their influence
existing law, the tax rate and the share of the local tax (“pre-
was probably of secondary importance.2 The dominant
cept”) to be paid to the council are to be determined by the
factors in the choice of this strategy were the political and
councils. The Chiefdom Councils are required to collect
economic interests of national politicians in the SLPP
this tax, presumably on behalf of the local councils, and to
government, who supported decentralization for two
hand over the precept to the local councils.5 Moreover, the
main reasons.3
identification of the chiefdoms as a lower unit of adminis-
First, the SLPP was the prime victim of prewar political
tration and of local councils as the highest political author-
economy: the dismantling of local councils went hand in
ity at the local level is not acceptable to the chiefs, creating
hand with the concentration of power in Freetown and the
continuing tensions between the two institutions. The vari-
consolidation of APC power in Sierra Leone (Reno 1995).
ous laws governing the chieftaincy have not been repealed.
The SLPP decision to decentralize in 2004 may have been
The delay in addressing this issue leaves open the possibility
part of a more general attempt to reconfigure political insti-
of manipulation by the national government to pursue a
tutions in a way that reduced the urban/central bias and the
“divide and rule” strategy (Robinson 2010; Acemoglou,
potential for a return to the prewar political economy,
Robinson, and Verdier 2004), thereby keeping the local
which tended to benefit the APC more than the SLPP.
councils weaker than they could be.
Second, the SLPP may have perceived decentralization
Several other laws are inconsistent with the Local Gov-
as a good opportunity to meet popular expectations,
ernment Act 2004 or duplicate its provisions. Four laws in
enhance the government’s legitimacy, and increase politi-
particular place more control in the hands of the ministries
cal support for the ruling party—all likely outcomes given
of education, health, and energy and power than envisaged
that a key source of popular discontent before the war was
in the Local Government Act 2004.6
the “local despotism” (Richards 1996) of the paramount
chiefs. The SLPP addressed that discontent by reestablish-
Political decentralization
ing the institution of local councils, which reduced the
authority of the paramount chiefs. The SLPP president, The Local Government Act of 2004 identifies local councils
Ahmad Tejan Kabbah, had been a district commissioner as the highest political authority in their jurisdiction. The
himself and often made public his view of local councils as legislation sets out a detailed political framework covering
cornerstones of democratic life and citizens’ participation the election and composition of councils; the qualifications
in local politics.4 of councilors; procedures for the election of mayors (urban)

CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 143
 and chairpersons (rural); powers to make and execute
bylaws; the role and responsibilities of ward committees;
and provision for citizen participation, transparency, and
accountability. The act also recognizes the laws and regula-
tions governing the chieftaincy and chiefdom administra-
tion, which were not repealed. Chiefdoms are identified
as the lowest unit of administration. The Local Government
Act provides paramount chiefs representation in councils
and membership in the ward committees.
Local elections in 2004 and 2008 were successfully
completed, and transitions were peaceful. Elections were
fairly competitive. In 2004, 1,112 total candidates regis-
tered with the National Electoral Commission for the 394
constituencies. In 84 constituencies (21 percent), coun-
cilors were elected unopposed. Elections in urban areas
were more competitive than in rural areas, and the elections
were more competitive in 2008 than in 2004, with the num-
ber of uncontested wards dropping from 84 to 38 (less than
10 percent of all constituencies). There was a significant
turnover of councilors between the two elections, creating a
new class of local politicians, perceived by the population as
“the young generation of leaders on the political scene”
(Zhou 2009, 105). These trends suggest that channels of
political accountability are taking root at the local level, with
citizens willing to reward or punish politicians based on
their performance.
Fiscal decentralization
The Local Government Act provides a framework for fiscal
decentralization for a first phase (2004–08).7 With the
exception of the point regarding local taxes noted earlier, it
clearly lays down the revenue domain of the local councils.
Three sources of financing exist for local governments in
Sierra Leone: central government transfers for devolved
functions and administrative expenses; local councils’
own revenues from taxes, fees, licenses royalties, mining
revenues, and other sources; and loans and grants from
other sources.8 The act recognizes that in the short term,
own revenues are unlikely to finance the functions
devolved to the subnational level. It therefore provides for
a “first-generation” system of intergovernmental transfers
to fill the gap.9
The act provides for tied grants to the local councils to
carry out the functions devolved to them and to meet their
administrative costs. The law separates the grant arrange-
ments for a first phase (2004–08) from those for the period
after 2008. The description of the goals and bases for grants
is different for the two periods. For the period up to 2008,
144 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
the grant for each devolved function was to be sufficient to
provide the service at the standard at which it was provided
before devolution. The act did not provide a rule or for-
mula for determining the vertical pool of resources to be
devolved; in practice, each sector allocation was deter-
mined through negotiations with the line ministry, inter-
mediated by the Local Government Finance Department
(LGFD).10 Allocation across councils was to be determined
on the principal of “equity.” Equity was not defined in the
law but, in practice, transparent formulas based on popula-
tion and existing infrastructure were devised and are being
used for the horizontal allocation of grants for devolved
functions. Additional administrative grants were provided
based on expenditure needs and fiscal capacity and indexed
to inflation.11
For the period after 2008, the law indicates that the vol-
ume of grants should allow councils to provide devolved
services “at an appropriate standard” and that the annual
changes must grow at least as fast as the total budgetary
appropriation made to government ministries. For all
grants for devolved functions, parliament retains the
authority to specify the functions on which these tied grants
must be spent. Horizontal distribution is expected to
depend on the expenditure needs and revenue-raising
capacity of local councils (the specific relationships are not
articulated).
Transfers are significantly tied not only to sectors but to
specific activities and programs, making for a large number
of separate grants. Although the formulas for the horizon-
tal allocation of grants are transparent, there is scope for
simplification and improvement. The design of a “second-
generation” grant system has been delayed and is currently
ongoing.
Budgeted transfers to local councils increased from Le
19 billion in 2005 to Le 34 billion in 2009, an increase of
44 percent (figure 8.2). With the exception of 2007
(a national election year), actual transfers increased during
the same period by 150 percent, reflecting the fact that
the ratio of actual to budgeted transfers rose from about
70 percent in 2005 to about 98 percent in 2009. The provi-
sional figures for 2010 indicate that the budgeted amount
for 2010 was more than 50 percent higher than for 2009
and that the actual transfers were higher than the budgeted
amounts. This is a remarkable achievement for a country
with a history of centralization. Transfers as a share of cen-
tral government’s nonsalary, noninterest expenditures also
increased, from 4.9 percent in 2005 to 11.9 percent in 2009
On average, however, actual transfers remain low, and local
councils are still underfunded relative to the functions that
Figure 8.2 Budgeted and Actual Transfers to Local Figure 8.3 Local Council Own Revenues as Share of
Councils, 2005–2010 Total Revenues, 2005–2010

55 54.86 35
33.3 32.9
50 51.42
30

Percentage
45
26.4
Billion leones

40 35.0 33.2 33.7

35 25 24.5 24.5
30 25.4
33.1 20
25 19.0 25.2 18.8
20 25.6
15
15 2005 2006 2007 2008 2009 2010
10 13.2
7.5
5 Source: IRCBP 2010b.
2005 2006 2007 2008 2009 2010 Note: Figures for 2010 are provisional.
Budgeted transfers Actual transfers

Source: IRCBP 2010b.

Note: Figures for 2010 are provisional.
have been devolved to them. Given these circumstances, the
move to a more rational “second generation” of grants is
long overdue.
Not surprisingly, revenue generation by the local coun-
cils continues to be weak, with local councils on average
able to finance only about 25–30 percent of their expendi-
tures from their own revenues (Searle 2009). Local taxes
and property taxes are the sources of revenues for the local
councils. The main sources of nontax revenues are market
dues, business registrations, license fees, and mining royal-
ties (in selected councils).
Local council revenues as a percentage of total revenues
indicate that limited and uneven improvements have taken
place over time (figure 8.3). This trend is slightly more pos-
itive if one considers revenues in absolute terms: in 2008 the
local councils collected Le 8.3 billion, a substantial increase
over the Le 4.7 billion collected in 2005. Revenues in 2009
were Le 16.2 billion, an increase of 95 percent over 2008.12
There was a further increase of about 10 percent between
2009 and 2010 although the share of own revenues in total
revenues fell due to the significant increase in the total vol-
ume of transfers.
Administrative decentralization
The statutory instrument accompanying the Local Govern-
ment Act 2004 provided a detailed framework and timetable
for devolving functions housed in 17 ministries, depart-
ments, and agencies to the local councils. The functions that
were to be devolved between 2005 and 2008 included pri-
mary and mid-secondary education, primary and secondary
health facilities, feeder roads, agriculture, rural water, solid
waste management, youth and sport activities, and some
fire and social welfare functions. Central ministries and
agencies retain responsibility for strategic planning, setting
of standards, quality control, and monitoring, as well as
procurement of certain priority commodities, such as text-
books and drugs. Central government also retains adminis-
trative control over staff responsible for performing
devolved functions.
The devolution of functions has been slower than
expected, with varying degrees of responsiveness. Although
the act envisioned completion of the transfer of authority
by 2008, by mid-2010 only 46 of 80 functions had been for-
mally devolved to local councils (Decentralization Secre-
tariat 2010). Important areas such as the devolution of
feeder roads remain politically contentious.
Each local council has a political head (mayor and deputy)
and administrative head (chief administrator) (figure 8.4).
The chief administrator is responsible for management and
all administrative and technical matters. In addition, each of
the local councils now has its complement of core technical
staff under the supervision of the chief administrator. How-
ever, staff for devolved functions continue to remain under
the administrative control of the central ministries, depart-
ments, and agencies. Under the new decentralization policy
(Government of Sierra Leone 2010a), staff for devolved
functions are to be fully devolved to the local councils by
2016. In the interim, it is proposed that administrative con-
trol of local councils over such staff will be enhanced
through “letters of deployment.”
MAJOR ACHIEVEMENTS OF
DECENTRALIZATION
Sierra Leone’s success in reestablishing local government in
a fragile postconflict environment is notable. Despite some

CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 145
 Figure 8.4 Intergovernmental Relationships under Decentralization

CENTRAL GOVERNMENT

Ministry of Finance Ministry of Local Government and

Line
Local Government Rural Development
ministries
Financial Department Decentralization Secretariat

Capacity
support
3 provinces
(deconcentrated units) Grants

19 local councils Staff for

District devolved functions devolved
officers (LGA 2004) functions

Share of
149 chiefdoms local taxes
paramount chiefs
• Local taxes
• Land

Source: Authors.
Note: The shaded boxes identify central government institutions and deconcentrated units (provinces, district officers). The dashed box identifies the key
actors of decentralized administration. The solid arrows indicate a clear relationship of authority across levels of government. The dotted arrows denote
the lack of clarity in defining relationships between government institutions, including the recently re-introduced district officers and local councils; local
councils and chiefdoms; and local councils and local staff, who are formally accountable to the local councils but appointed by the central administration.
The hollow downward-slanting arrow indicates the flows of resources from the central government to local councils (resources come from capacity sup-
port from the Ministry of Local Governance and grants from the Ministry of Finance). The hollow upward-slanting arrow indicates the flows of resources
(share of local taxes) from chiefdoms to local councils. Historically, there were 149 chiefdoms established under the Chiefdoms Councils Ordinance 1938
(and subsequent amendments). A local council thus has more than one chiefdom within its jurisdictions.

of the shortcomings outlined above, decentralization is
now well established, arguably more strongly on the politi-
cal dimension than on the others. Two council elections
have been completed; all local councils have the core staff
to carry out planning, budgeting, accounting, and procure-
ment functions; a system of intergovernmental transfers is
in place; and, although there is scope for improvement,
local governments are able to work with centrally managed
frontline staff to manage service delivery in the areas
devolved to them. The pace at which local councils assumed
full identity as democratic, effective, and legitimate political
actors and institutionally oriented themselves to discharg-
ing devolved functions has been a notable feature of the
decentralization process.

146 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
Service delivery
With all councils decentralizing at the same time, it is dif-
ficult to rigorously establish the extent to which improve-
ment in services is a result of decentralization.13 That
said, on average the quality of services did not decline,
and in some cases there have been significant improve-
ments, according to data from a series of national public
service and sector (health and education) surveys carried
out by the evaluation unit of the Institutional Reform and
Capacity Building Project Evaluation Unit (IRCBP)
together with the Abdul Latif Jameel Poverty Action Lab
(table 8.1).14
Communities far from Freetown but close to a district
capital saw the biggest improvements in services, even
holding constant the remoteness of these communities. A
reduction in “distance from power”—a direct byproduct of
decentralization—thus appears to have had a positive
impact on service delivery.
These improvements are remarkable if one considers the
short time since the launch of decentralization and the fact
that local government performance continues to be con-
strained by several factors, including the incomplete devo-
lution of functions and line staff, the relatively small size
and tied nature of the transfers, and the unresolved ten-
sions between local councils and the traditional authorities.
Health. Health is the sector that has progressed furthest
on devolution, with about $3 million, just less than one
quarter of the national health budget, budgeted to grants to
local councils as early as 2006. Access to and quality of
health services have improved dramatically since 2005,
with most of the gains taking place between 2005 and 2006.
Clinic infrastructures, availability of drugs, and numbers of
staff have all improved, with the result that public satisfac-
tion with health services improved from 81.0 percent to
Table 8.1 Quality of Service Delivery Reported in Household Surveys, Selected Years
(percent of respondents)
Service indicator
Access to school within 30-minute walking distance
Satisfaction with primary schools
Satisfaction with health clinic
Spoke to an extension worker in past year
Access to sufficient storage space (farming households only)
Drivable road within 30-minute walking distance
Market area within 60 minutes
Water source within 15 minutes
Source: IRCBP 2010.
CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
90.6 percent (see table 8.1). Between 2006 and 2008 the
largest gains were in the number of staff, particularly sen-
ior staff. Although progress is being made in filling vacant
positions, the percentage of clinics open fell between 2006
and 2008, from 88 percent in 2006 to 82 percent. Also, the
percentage of clinics receiving supervision by Ministry of
Health and local council members declined after 2005.
Despite decentralization, local councilors in particular do
not appear to be taking a strong supervisory role, with only
one in four clinics receiving a visit from a councilor in the
year preceding the survey. These trends suggest that more
effective supervision by the local councils and ministry
staff, rather than more hiring, should be the main focus
moving forward.
Education. Education has seen less devolution than other
sectors. Because only one detailed survey of school quality
has been undertaken by the IRCBP, it is harder to assess
gains in education. Household surveys show that more
rural Sierra Leoneans are within reach of a primary school
than they were in 2005. The percentage of households with
access to a school within 30 minutes’ walking distance
increased from 68.3 percent in 2005 to 74.3 percent in 2008
(see table 8.1). Overall public satisfaction with primary
schools has improved. Informal school fees remain high,
however, a key reason why children are not in school.
Although only 3 percent of schools were reported to be
closed, teacher absenteeism was estimated at 22 percent in
2005, highlighting the lack of adequate supervision.
Agriculture. Access to drying floor and storage space
improved sharply between 2005 and 2007, although most
households still lack access. Only 18 percent of households
had contact with an extension worker in 2007, down from
23 percent in 2005. Nearly half of farmers sell their produce
2005 2007 2008
68.3 73.9 74.3
87.7 94.4 90.3
81.0 90.9 90.6
23.0 17.8 9.0
8.4 11.8 14.3
67.1 73.2 77.5
31.9 45.8 50.9
61.0 73.4 80.9
147
 to traders who come to the village, a third sell at the market
themselves, and the rest sell to a trader at market.
Other sectors. Although the Local Government Act was
supposed to devolve responsibility for the rehabilitation
and maintenance of roads to local councils, devolution has
not yet taken place, and no money has been transferred
from the central government to the local councils for this
function. Resistance by the powerful Sierra Leone Roads
Authority is the main reason for this delay.15
No central government agency is responsible for mar-
kets, another area in which local councils do not yet receive
transfers. Local councils have, however, spent a significant
share of their discretionary funding on roads and markets,
which are perceived as key means to increasing the effi-
ciency of the agriculture sector.
Access to markets in rural areas remains weak, with
half of the rural population having to travel more than an
hour to reach a market. As indicated in table 8.1, however,
there have been significant and steady improvements: in
2005 only 32 percent of rural respondents had a market
less than 60 minutes away. In 2008 this figure increased to
51 percent. Road access also improved, but the percentage
of respondents having access to regular public transport
declined between 2005 and 2008. Access to water and
water sources saw significant improvements: in 2008
about 81 percent of respondents had access to a water
source within 15 minutes, a sharp increase from the 2005
baseline (61 percent).
Local governance
A central argument in support of decentralization is that it
brings government closer to the people. By expanding the
political space, decentralization is expected to allow greater
government accountability and citizen participation to pub-
lic affairs, hence strengthening state legitimacy (Bardhan
and Mookherjee 2006).
In Sierra Leone national elections tend to become “win-
ner takes all” games focused on region-based ethnic identi-
ties. The role of ethnicity is likely to be less relevant at the
local level, where voters are more likely to share a common
ethnic background. Decentralization may provide citizens
with greater opportunities to monitor the performance of
local authorities on service delivery and hold them account-
able through elections. The experience of Sierra Leone sug-
gests that decentralization is indeed contributing to better
local governance by providing greater scope for citizens’ par-
ticipation and engagement with local authorities, especially
148 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
in rural communities and areas without transport or
telecommunications (IRCBP 2010).
Accountability and participation. Provisions of the
Local Government Act stipulate that councilors must
declare their assets and councils must maintain an inven-
tory of assets, print receipts, and maintain other accounting
documents. Notice boards are also to be maintained in
wards and at the council displaying financial information
and strategic documents, such as development plans and
procurement contracts, to enhance accountability and
information sharing with the community. To ensure greater
accountability and support a participatory approach to the
decision-making process, ward committee members partic-
ipate in council meetings and review the council’s develop-
ment plan. It is the responsibility of the ward committee to
prioritize the community’s development needs, which are
then finalized and passed at the council meeting.
Evidence suggests that local communities are becoming
increasingly active in demanding services and holding local
authorities accountable, looking at local elections as a means
to achieve the promise of development.16 Participation in
local elections remains moderately high—albeit lower than
in national elections—with civic activism tending to be
higher in remote areas than in urban areas.17
Marginalized groups, such as women and ethnic
minorities, have been the largest beneficiaries of the new
space for political participation. In 2004 women occupied
about 13 percent of council seats in council elections; in
2008 this share increased to 18 percent, more than three
times the 5 percent share of seats that women have in the
national Parliament (IRCBP 2010). In 2004 and 2008 repre-
sentatives from minority ethnic groups such as the Kono,
Loko, and Sherbro, were elected to local councils (Zhou
2009)—a remarkable achievement given Sierra Leone’s his-
tory. Decentralization has thus made significant contribu-
tions to promoting prospects of political stability by allowing
greater participation and power-sharing dynamics.
Performance of local councils. In 2006 the Decen-
tralization Secretariat, supported by the IRCBP, designed
a Comprehensive Local Government Performance Assess-
ment System (CLoGPAS) tool to serve as a sustainable
local council management accountability mechanism for
the local councils.18 The first assessment was carried out
in June 2006. A follow-up assessment was conducted in
early 2008. The results show that performance of local
councils is improving in several development and man-
agement functions.
Figure 8.5 Number of Local Councils Meeting Minimum councils making the largest gains among political institu-
Conditions, 2006 and 2008 tions between 2007 and 2008 (table 8.2). Together with chief-
dom officials, local (political) authorities are more trusted
18 by citizens than national government officials. These
16
14
trends are encouraging, although some of the findings may
12 partly reflect a temporary boost from the publicity sur-
Number

10 rounding the July 2008 local council elections. This line of

8 argument seems to be (indirectly) confirmed by the findings
6
4
of the 2008 National Public Services survey that trust in local
2 councils is higher among citizens who are more actively
0 engaged with local politics.
t

ng

io ,

ity
nc luat ion
en

en
tin

nc
ni

ac
na n
em

em

re
un

t
lan

ev nta

ap
pa
co
ag

ur

lc
tp

nd me
ns
an

oc
ac

THREATS AND CHALLENGES

en

a
ra

g a ple
lm

Pr
nd

tio
pm

rin im
cia

ga
elo

Fu
an

in

ito ct

Six years after the decentralization initiative was launched, it

ev

et

on je
Fin

m Pro
dg
D

has arguably taken firm root, leading several observers to

Bu

suggest that it would be very hard to reverse the process and

2006 2008 concentrate power in Freetown again. Thus, from a
Source: Decentralization Secretariat 2007, 2009. Freetown-centric perspective, decentralization is a “genie”
that is out of the bottle. Recent developments, however,
suggest that pressures might be building to weaken decen-
Figure 8.5 shows the number of councils that fully met tralization, casting some doubt on the central government’s
the seven “minimum conditions” of the CLoGPAS.19 It intentions to unequivocally move the agenda forward.
indicates that the number of local councils that met the Ironically, this apparent weakening may well be a result of
minimum conditions rose between 2006 and 2008 in all but the successes of the initiative. Rather than empowering
two areas (transparency and, to a lesser extent, functional local councils further, national elites seem to be following a
capacity). Local councils hold regular meetings and pro- strategy of “divide and rule,” preventing local councils from
duce minutes, citizens participate in meetings, participa- becoming strong enough to seriously challenge the political
tory development planning has taken root (although there hegemony of the center.
is scope for improving its quality), accounts are completed
on time, and regularly audited and financial information is Stance of the national government
disclosed. These promising trends have been confirmed by
the latest (draft) report on public expenditure and financial Tensions between the local councils and the chieftaincy
accountability (Government of Sierra Leone 2010b), which result largely from a lack of clarity about their respective
assesses the performance of central and subnational gov- domains (Fanthorpe 2005; Sawyer 2008). It is difficult to
ernment authorities across various dimensions of public judge whether this situation is intentional or reflects simple
financial management.20 According to the report, in 2010
local councils received the highest scores on key dimen-
Table 8.2 Citizens’ Trust in Public Institutions,
sions of budget execution (namely, competition, value for 2007 and 2008
money, and controls in procurement) and accounting prac- (percent of respondents)
tices (namely, timeliness and regularity of accounts recon-
Type of official 2007 2008
ciliation), scoring higher on these dimensions than the
Justice sector
national government. This performance is remarkable in Police 36 38
view of the fact that local councils were established as Local court 41 45
Magistrate court 34 37
recently as 2004.
Political authorities
Central government 42 44
Citizens’ trust in public institutions. Results from Chiefdom 47 51
National Public Services surveys indicate that public confi- Local council 33 49

dence in local councilors has increased over time, with local Source: IRCBP 2010.

CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 149
 delays in decision making, although some evidence suggests
that it is intentional. A national decentralization policy has
recently been drafted, but a critical companion piece—the
policy on chiefdom governance—has not been prepared.
The Chieftaincy Act of 2009 codifies and adds to customary
law on the election and removal of chiefs. With the enact-
ment of this law, any reforms aimed at the democratization
of the chieftaincy are unlikely, at least in the medium term.
Under the law, the central executive has effective leverage
over the chieftaincy, because it plays an important role in
the election and removal of paramount chiefs. The ministry
responsible for local government has traditionally dealt with
the chiefs and even today is much more comfortable with
this function than with its new role with respect to the local
councils.21 These links are important for the national elec-
tions because, by some estimates, the chiefs are able to influ-
ence 10–20 percent of voters in their jurisdictions.
In 2009, contrary to the provisions of the Local Govern-
ment Act 2004, the minister for local government advised the
chiefs not to share local tax revenues with the local councils.
Since then, also contrary to the provisions of the law, the
ministry has played a role in determining the amount of the
local tax precept, establishing a range of 0–20 percent for
different classes of chiefdoms. Until 2008 most chiefs were
paying a precept of 60 percent. This ministry action will
undermine the fiscal autonomy of the local councils.
The National Decentralization Policy approved by the
cabinet in September 2010 provides strong support to the
hypothesis of a containment strategy by the center to limit
political power of local governments. According to the pol-
icy, “Local councils shall continue to exist as the highest
development and service delivery authority” (Government
of Sierra Leone 2010a). This policy is inconsistent with the
Local Government Act (2004), which defines local councils
as the highest political authority at the local level. The pol-
icy also reintroduces the position of district officers.
In June 2010 the APC government decided to reintro-
duce the post of district officers, who traditionally repre-
sented the national government in the districts and
provided the links with the chieftaincy. The official explana-
tion is that there is a need to establish a stronger channel of
communication between the national government and the
chiefs. Representatives from the SLPP opposition and civil
society groups, however, have articulated the concern that
the district officers—acting as representatives of the central
government at the local level—will try to influence decision
making by the local councils, which may further reduce the
councils’ financial autonomy and increase the control of the
ruling party as the 2012 elections draw closer. Several
150 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
senior government functionaries have remarked that the
local councils—whose key function is development—were
becoming political. Although these actions by the central
government may be a response to real political change
engendered by decentralization, they represent a potential
threat to the autonomy of the local councils and, more
generally, to prospects for further strengthening decen-
tralization in Sierra Leone. The incentives of members of
parliament (MPs) are aligned with those of the central
executive. The national election of 2002 was based on
proportional representation and the MPs did not have
geographical constituencies. With the change in 2007,
MPs represent single-member constituencies and now see
the local councils as competitors.
Champions of change
The Decentralization Secretariat (Dec Sec) and the Local
Government Finance Department (LGFD) have played crit-
ical roles in implementation and have been strong advocates
for decentralization. Both are extragovernmental agencies
created under the Institutional Reform and Capacity
Building Project financed by the International Development
Association (IDA) and a multidonor trust fund financed by
the European Commission and the U.K. Department for
International Development. Although, in principle the Dec
Sec is an arm of the Ministry of Local Government and Rural
Development and the LGFD is a unit in the Ministry of
Finance and the secretariat to the Local Government Finance
Committee (LGFC), all of the staff and expenses associated
with these agencies are currently financed by the IRCBP.22
With the project due to close soon, it is not entirely clear
how and how well their functions will be integrated into
government. Without the mainstreaming of these functions,
the future of decentralization in Sierra Leone will be in jeop-
ardy. If progress on this issue of mainstreaming continues to
be slow, donors will need to make a judgment fairly soon
about whether the government is genuinely handicapped by
financial and capacity constraints or whether this lack of
progress reflects apathy toward the decentralization agenda.
Dependence on national government
The local councils are highly dependent on grants from the
national government. Although the volume of these grants
has increased over time and a larger proportion of budgeted
amounts are now being transferred, the lack of financial
autonomy leaves the local councils very vulnerable to the
whims of the national government. The lack of control over
frontline staff weakens the ability of the local councils to
influence the quality of services. Both of these risks are exac-
erbated by the fact that the national government appears to
want to keep the local councils weak.
LESSONS LEARNED AND CONCLUDING
REMARKS
Decentralization has had a significant positive impact on the
political landscape in Sierra Leone. The sharing of political
power beyond Freetown is unprecedented; on this dimen-
sion alone, the initiative can be viewed as a major success.
There is scope for improving the capability of local coun-
cils to deliver on development outcomes. The volume of
resources available to the local councils remains small, and
the relatively low level of own revenues and dependence on
central grants undermines accountability, limits auton-
omy, and makes the local councils vulnerable to manipu-
lation by the center.
Functioning and effective local governments provide a
viable and promising alternative to the persistent patrimo-
nial nature of Sierra Leone’s national politics for a shift
toward public goods and better service delivery to the
poor.23 In particular, the local councils are in a position to
contribute to the improvement of services and the provi-
sion of public goods in the social sectors and in those
aspects of infrastructure (water and sanitation, feeder
roads, off-grid power) that can efficiently be provided at
the subnational level.
Although it is unlikely that any government will openly
seek to recentralize power in Freetown, as the APC led by
Siaka Stevens did in 1972, there are risks that the center will
try to undermine the emerging political power at the sub-
national level. The reintroduction of the district officers
and the downgrading of the political status of the local
councils provide some evidence of a new containment strat-
egy designed to protect the interests of the national elites.
These efforts by the national government to undermine the
autonomy and strength of local councils may be the ulti-
mate outcome of an effective decentralization process—the
“paradox of success”—implicitly suggesting that local gov-
ernments are indeed emerging as important players in the
political and development landscape of Sierra Leone. The
risks associated with the containment strategy of the center
suggests that donors and civil society groups will need to
play an important role, however, if decentralization is to suc-
ceed and central government efforts to undermine it are to
be neutralized. Continuing engagement by donors and key
national stakeholders is crucial to support decentralization
CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
and help local authorities in their ongoing efforts to meet
citizens’ demands and perform the functions devolved to
them. Donors have played an important role so far. The
main instrument has been the IRCBP, which, through the
Dec Sec and LGFD, has supported the establishment of a
simple but robust intergovernmental grant system; made
resources available for small capital works; and, probably
most important, supported the development of capacity in
the local councils to enable them to perform their core
functions and become established as credible governments.
On some dimensions, such as procurement, capacity is
rated to be better at the subnational level than at the
national level. Continuing support for strengthening ser-
vice provision at the local level and enhancing the resources
and autonomy of the local councils is being provided
through the $20 million IDA–financed Decentralized Ser-
vices Delivery Project (DSDP). This second-generation
project—which tilts the balance toward significantly aug-
menting the grants provided by the central government
while providing support for a “maintenance” level of capac-
ity development—is expected to attract additional donor
contributions. An increasing allocation of donor resources
to the subnational level through well-designed projects will
strengthen both the autonomy and the capacity of local
councils and help mitigate countervailing pressures from
the national government. In the authors’ view, there is a
continuing need for building technical capacity beyond
what was originally envisaged in the DSDP and, more
important, for dialogue and advocacy.
These donor initiatives will need to be complemented
with additional reform efforts to improve the prospects of
decentralization and create the conditions for a greater
impact on service delivery:
■ A second-generation system of grants needs to be
designed and implemented as soon as possible to
enhance predictability and autonomy and provide incen-
tives for efficiency. The divisible grant pool needs to be
determined based on a set of well-understood forward-
looking criteria, the allocation across councils should be
based on an agreed philosophy for equalization, and
grants need to be gradually untied to provide greater dis-
cretion to the local councils.
■ Local councils need to be gradually provided with greater
administrative control over frontline staff associated
with service delivery. As noted, the provisions of the new
national decentralization policy identify a vision for
2016, when comprehensive devolution of staff to the
local councils should have taken place and interim
151
 arrangements made during which authority is gradually
handed over to the local councils. While this is, in prin-
ciple, a sensible approach the transfer of responsibility to
local councils will have to be carefully balanced with
local council capacity to undertake this responsibility.
■ Issues concerning the functions and jurisdictions of sub-
national institutions need to be resolved to ensure that
local councils and paramount chiefs work better
together.24 This strategy would enable subnational polit-
ical actors to present a more united front, which would
be effective in strengthening the decentralization process
while mitigating the risk of manipulation from the cen-
ter. The national government needs to take the lead in
clarifying functions and jurisdictions.
■ Across all devolved functions, improved supervision and
monitoring of service delivery by local councils and cen-
tral ministries is needed to ensure that implementation
meets required standards.
■ As more resources become available at the subnational
level, the risk of rent-seeking behavior and corruption
will increase. Close watch will need to be kept to contain
opportunities for corruption and elite capture; additional
resources and autonomy will have to be complemented
with robust accountability arrangements involving the
ward committees and citizens’ participation.
Key country stakeholders and donors took advantage of
the window of opportunity that became available after 2002
by supporting the enactment and implementation of the
Local Government Act. They did so knowing that it was not
a perfect document and that several loose ends needed to be
resolved. Their actions helped to establish a system that is
reasonably robust and to create a new class of subnational
actors that is gaining central prominence in shaping new
political and economic dynamics. The genie is now out of
the bottle. The question is whether it will be able to work its
magic or whether it will be “contained” by the center.
Donors, together with civil society and other stakeholders,
will have to play a critical role in maintaining momentum
along the trajectory that has been established, including
advocacy initiatives and continuous policy dialogue with
the government. Any letup in effort at this stage could jeop-
ardize the initiative and compromise the remarkable
progress achieved since 2004.
NOTES
1. This massive bias not only created large socially unde-
sirable biases in resource allocation, it also led to political
152 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
instability, by raising the stakes of politics, making the cen-
ter attractive to capture in a zero-sum game.
2. The standard arguments about the benefits of decen-
tralization focus on the efficiency of service delivery. The
basic idea is that by bringing representatives closer to the
people, decentralization leads to socially more desirable
service provision, because local politicians are more
accountable and because they have better information
about people’s preferences and what needs to be done.
Arguments that link decentralization to better service deliv-
ery may be particularly compelling in situations in which
there are important ethnoregional divisions in a society,
making it difficult to reach agreement about which policies
to adopt or public goods to supply (Robinson 2010; see also
Bardhan 2002 and Eaton, Kaiser, and Smoke 2010 for a
review of this argument).
3. An additional factor that may have driven the SLPP to
decentralize can be found in the use of decentralization as
an electoral strategy to ensure political survival at the
local level in the event of a loss of power at the national
elections. Although theoretically plausible and historically
relevant in other contexts (O’Neill 2003), this kind of
electoral incentive was hardly a dominant factor in Sierra
Leone. Historical evidence suggests that at the time decen-
tralization was launched (2004), the SLPP was confident
of remaining in power and that it remained confident of
doing so until early 2007, when the sudden realization of
possible electoral loss prompted uncoordinated and inef-
fective actions (see Kandeh 2008 for an excellent analysis of
the 2007 elections).
4. At the launch of the Local Government Reform and
Decentralization Programme, on February 20, 2004, Presi-
dent Kabbah remarked, “People have the right to elect the
leaders, men or women, who are supposed to serve them at
the national level, in Parliament. They also have the right to
choose those who should serve them at the district and
other local levels. This, in my view, is what the principle of
democratic decentralization is all about. It should and must
remain a cornerstone of the process of nation-building in
the country.”
5. The Local Taxes Act of 1975 identified the “local” tax as
a tax to be collected by the “local authority” and paid into
the Consolidate Revenue Fund/Accountant General. Under
this act the rate of the tax was to be determined by the min-
ister for internal affairs. Under this formulation, the local
tax is collected by the “local authority” on behalf of the
national government. Section 45 (4) of the Local Govern-
ment Act 2004 made the “precept” payable to the local coun-
cils but did not make any reference to the Local Taxes Act
1975, which is still on the statute books.
6. The laws are the Education Act, the Hospitals Board Act,
the SALWACO Act, and the Local Tax Act.
 7. See Fox 2009 and Searle 2009 for detailed discussions of
fiscal decentralization in theory and practice.
8. Under section 65 of the Local Government Act 2004 and
section 17 of the Public Debt Management Bill 2010
(gazetted but not enacted at the time of writing), local gov-
ernments can borrow domestically up to limits agreed with
the Minister of Finance. This borrowing need not be guar-
anteed by the central government. So far, local governments
have not used this option as a source of financing and local
government debt is not currently a problem.
9. We refer to the grant arrangements envisaged under the
Local Government Act 2004 for the period 2004–2008 as the
“first generation” system of intergovernmental transfers.
The design of a “second generation” system is ongoing and
has not been completed.
10. The LGFD is a unit in the Ministry of Finance that per-
forms the secretariat functions for the Local Government
Finance Committee (LGFC), the entity that determines the
volume and distribution of grants under the law.
11. In practice the administrative grants are in two parts.
The first part covers the sitting fee and transport allowances
of councilors and is based on the number of councilors. The
second part finances their general administrative expenses;
it is positively related to expenditure needs for undertaking
revenue collection and administering devolved functions
and negatively related to fiscal capacity.
12. Several factors (beyond decentralization) may have
driven the positive results in 2009; considered alone, this
sharp improvement is not sufficient to establish a trend.
13. For example, the fact that satisfaction with health
improved more sharply than satisfaction with education
does not necessarily reflect the fact that health decentralized
earlier than education. It may be that health decentralized
more rapidly than education because there was greater
capacity in the Ministry of Health and that satisfaction with
health improved more because of that capacity.
14. The National Public Services survey was administered
three times between 2005 and 2008. The quality of services
at health clinics was surveyed three times, and a baseline for
education quality was created in 2005. The surveys include
large, nationally representative samples of households, clin-
ics, and primary schools; the data collected can be com-
pared over time.
15. The Sierra Leone Roads Authority is a powerful inde-
pendent statutory body established in 1993 responsible for
policy formulation and implementation in the roads sector.
16. A household survey by GoBifo (Sierra Leone’s
community-driven development project) and the IRCBP in
2009 in the Bonthe and Bombali districts asked respondents
their main reasons for voting. It found that in national elec-
tions, voters ranked political party as the most important
CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
and development second; for local elections the promise of
development was the priority (Zhou 2009).
17. According to the latest National Public Services survey
data, respondents were significantly more likely to report
voting in national (87 percent) than local council (77 per-
cent) elections. Official electoral data show substantially
lower participation rates in local elections, indicating a
decrease from 55 percent in 2004 to 39 percent in 2008.
18. The design and implementation of CLoGPAS involves
the setting up of a multidisciplinary task team comprising
technical staff of IRCBP and the Ministry of Internal Affairs,
Local Government and Rural Development.
19. The minimum conditions deal with aspects of local
council management accountability and examine func-
tional capacities of the local councils in terms of their
preparedness to take over devolved functions and deliver
services at acceptable standards. They also assess/review the
compliance of local councils with existing laws and regula-
tions that guide the decentralization process.
20. More precisely, the public expenditure and financial
accountability report “first examines the credibility of the
Budget as a tool for implementing government policy, and
then looks at two key crosscutting issues relating to Public
Financial Management (PFM), the comprehensiveness and
transparency of PFM systems. It then rates performance
through the four key stages in the budget cycle: budget for-
mulation, budget execution, accounting and reporting and
finally external scrutiny and audit. Under each dimension, a
set of performance indicators is identified, and scoring cri-
teria is set out” (Government of Sierra Leone 2010b).
21. In 2004 the ministry was known as the Ministry of
Local Government and Community Development. In 2007
it became the Ministry of Internal Affairs, Local Govern-
ment and Rural Development. In November 2010, Internal
Affairs was hived off, so the ministry is now the Ministry
of Local Government and Rural Development.
22. The IRCB Project was launched in mid-2004, with a
World Bank credit of $25.1 million. Financing was
extended in 2006 with an additional $25 million provided
by the Department for International Development and the
European Union through a multidonor trust fund.
23. See Robinson (2008) for a compelling analysis of the
sources of patrimonialism in Sierra Leone. Robinson
(2010) further elaborates on the political economy of
decentralization, analyzing the reasons why local govern-
ments are expected to be less subject to the capture of pat-
rimonial politics.
24. Fanthorpe and Sesay (2009) make a number of well-
informed and constructive recommendations for how to
reform the chieftaincy and make it work better with local
councils.
153
 REFERENCES
Acemoglu, Daron, James A. Robinson, and Thierry Verdier.
2004. “Kleptocracy and Divide-and-Rule: A Model of
Personal Rule.” Journal of the European Economic Associ-
ation 2 (2–3): 162–92.
Bardhan, Pranab K. 2002. “Decentralization of Governance
and Development.” Journal of Economic Perspectives 16
(4): 185–205.
Bardhan, Pranab, and Dilip Mookherjee, eds. 2006. Decen-
tralization and Local Governance in Developing Countries:
A Comparative Perspective. Cambridge, MA: MIT Press.
Decentralization Secretariat. 2007. Comprehensive Local
Government Performance Assessment System (CLoGPAS):
2006 Summary Report. Freetown.
———. 2009. Comprehensive Local Government Perfor-
mance Assessment System (CLoGPAS): 2008 Summary
Report. Freetown.
———. 2010. Status Report on the Devolution of Functions
to the Inter-Ministerial Committee. Freetown.
Eaton, Kent, Kai Kaiser, and Paul Smoke. 2010. The Political
Economy of Decentralization Reforms in Developing
Countries: A Development Partner Perspective. World
Bank, PRMPS, Washington, DC.
Fanthorpe, Richard. 2005. “On the Limits of the Liberal
Peace: Chiefs and Democratic Decentralization in Post-
War Sierra Leone.”African Affairs 105 (418): 27–49.
Fanthorpe, Richard, and Mohamed Gibril Sesay. 2009.
Reform Is Not against Tradition: Making Chieftaincy Rev-
elation in 21st Century Sierra Leone. Campaign for Good
Governance, Freetown.
Fanthorpe, Richard, Taylor Brown, Janet Gardener, Lansana
Gberie, and M. Gibril Sesay. 2006. Sierra Leone: Drivers of
Change. IDLgroup, London.
Fox, William F. 2009. “Devolution and Grant Structures for
Local Councils in Sierra Leone.” World Bank, AFTPR,
Washington, DC.
Government of Sierra Leone. 2010a. National Decentraliza-
tion Policy. September. Freetown.
_____ . 2010b. Public Expenditure and Financial Accounta-
bility (PEFA) Sub-National Government Summary
Report. Freetown.
154 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE
Hanlon, Joseph. 2005. “Is the International Community
Helping to Recreate the Preconditions for War in Sierra
Leone?” The Round Table: The Commonwealth Journal of
International Affairs 94 (381): 459–72.
IRCBP (Institutional Reform and Capacity Building Project
Evaluation Unit). 2010. Report on the IRCBP 2008
National Public Services Survey: Public Services, Gover-
nance, and Social Dynamics. Freetown.
IRCBP (Institutional Reform and Capacity Building Project
Evaluation Unit), and J-PAL (Abdul Latif Jameel Poverty
Action Lab). 2009. Basic Services and Decentralization in
Sierra Leone: Trends and Lessons. Abdul Latif Jameel
Poverty Action Lab, Massachusetts Institute of Technol-
ogy, Cambridge, MA.
Kandeh, Jimmy. 2008. “Rogue Incumbents, Donor Assis-
tance and Sierra Leone’s Second Post-Conflict Elections
of 2007.” Journal of Modern African Studies 46: 603–35.
Kieh, George. 2005. “State-Building in Post-Civil War Sierra
Leone.”African and Asian Studies 4 (1–2): 1.
O’Neill, Kathleen. 2003. “Decentralization as an Electoral
Strategy.” Comparative Political Studies 36 (9): 1068–91.
Reno, William. 1995. Corruption and State Politics in Sierra
Leone. New York: Cambridge University Press.
Richards, Paul. 1996. Fighting for the Rainforest: War, Youth
and Resources in Sierra Leone. Oxford: James Currey.
Robinson, James. 2008. Governance and Political Economy
Constraints to World Bank CAS Priorities in Sierra Leone.
World Bank, AFTPR, Washington DC.
———. 2010. The Political Economy of Decentralization in
Sierra Leone. World Bank, AFTPR, Washington, DC.
Sawyer, Edward. 2008. “Remove or Reform? A Case for.
Restructuring. Chiefdom Governance in Post-Conflict.”
African Affairs 107: 387–403.
Searle, Bob. 2009. Sierra Leone: Issues in Fiscal Decentraliza-
tion. World Bank, AFTPR, Washington, DC.
Truth and Reconciliation Commission. 2004. Final Report of
the Truth and Reconciliation Commission of Sierra Leone.
Freetown.
Zhou, Yongmei, ed. 2009. Decentralization, Democracy and
Development: Recent Experiences from Sierra Leone.
World Bank, AFTPR, Washington, DC.
 CHAPTER 9
Transport Infrastructure and the Road
to Statehood in Somaliland
Jean-Paul Azam
ow can economic success emerge in the midst of
H
political chaos and civil war? This is the question
raised by the experience of Berbera, a successful
port located on the southern shore of the Gulf of Aden, in
the former British protectorate of Somaliland. Berbera’s
traffic has been rising steadily over the past decade. For
example its traffic of import containers has nearly doubled
between 2003 and 2007 (MNPC 2010). This port’s promis-
ing position is confirmed by the fact that Bolloré Africa
Logistics, the biggest port operator in Africa, has recently
announced its plan to invest massively in Berbera’s port
and its transport corridor with the Ethiopian capital city,
Addis Ababa.1
Somaliland’s small population, estimated at about
3.5 million people, benefits from two main kinds of physical
assets: its pastoral assets of grazing land and livestock, on the
one hand, and its transport infrastructure, on the other
hand. The latter includes mainly the port of Berbera and, to
a lesser extent, the airport of Hargeysa, its capital city, as well
as their connecting roads. The nomadic herdsmen need the
port of Berbera for exporting their livestock to the Arabian
Peninsula and other Middle Eastern countries, which pro-
vide their main outlet. Moreover, Berbera is ideally located
to give a convenient access to the Indian Ocean to neigh-
boring landlocked Ethiopia, and there is a main road
between the two countries.2
Ethiopia’s Ogaden province across the border is also
peopled by Somali-speaking herdsmen, and that border is
traditionally fairly porous, as might be expected in a land
mostly devoted to nomadic pastoral activity. Refugees have
crossed this border back and forth over the past decades,
depending on the changing intensity of fighting during the
Ethiopian civil war, which ended in 1991, and the ongoing
one in Somalia that began in 1991. The best grazing land,
called the Haud, straddles this border, which nomadic
herdsmen also routinely cross (Doornbos 1993, map 6.1,
101). While ethnically homogenous, these Somali-speaking
people of Djibouti, Ethiopia, Kenya, Somalia, and Soma-
liland are traditionally affiliated with different clans that
straddle these countries’ borders and that at times enter
into violent conflict with one another. The chaos that pre-
vails in most of Somalia, which has been engulfed in war-
lordism and banditry ever since the collapse of the state in
1991, testifies to the threat of violence that looms over these
people. Siyad Barre’s military government had launched an
unsuccessful invasion of Ethiopia’s Ogaden region in 1978
and eventually signed a peace agreement with the Mengistu
regime in 1988, which was perceived as a disgrace by many
Somalians. Many military officers from the north, mainly
from the Isaaq clan, which is the largest one in Somaliland,
accused Siyad Barre of having mishandled the operation.
They felt that the conquest of a more limited area in the
Ogaden, including mainly the grazing land of the Isaaq
clan, could have been a success. The peace agreement with
Ethiopia triggered a full-blown civil war in Somalia,
because many clans wanted to hang on to the pan-Somali
155
 project, and a coup d’état toppled the military regime in
1991. Chaos ensued, and Somalia collapsed into a stateless
entity. Nevertheless, in the nearly two decades since
Somaliland seceded from Somalia, it has managed to
develop the port of Berbera and to enforce an acceptable
enough level of security on the paved road linking it to
Ethiopia for the traders to adopt this route. A large share
of Ethiopia’s international trade is now shipped through
this port. This achievement stands in sharp contrast to the
chaotic remainder of Somalia.
This chapter presents a framework for analyzing this
unexpected success and draws some lessons for under-
standing state formation in general. A very simple model
is presented that shows how the political equilibrium that
emerged in Somaliland is fundamentally rooted in the
need to provide security to the traders who provide most
of Berbera’s activity. This is a sine qua non for the traffic
through Berbera to be active and flourishing. This model
sheds some light on the fairly unusual political institutions
that emerged from less than two decades of self-rule by
Somaliland. Although some limited fighting occurred for a
while in the early 1990s, this former British colony quickly
engaged in a political process that led to the creation of a
fairly successful democracy in about a decade.
What seems very important when looking at the experi-
ence of Somaliland is that this gradual buildup of a function-
ing state started from the grassroots, with very little outside
interference. Eubank (2010) emphasizes that the Somaliland
Republic has not been recognized internationally, which
makes it ineligible for foreign aid. In his view, this is an asset
rather than a liability, because it forced the Somalilanders to
develop accountable political institutions and to engage in
state formation in a non-Eurocentric fashion. The aim of this
chapter is to go one step further in analyzing the type of gov-
ernment that resulted and in particular to explain clearly why
the “fiscal decentralization” that Eubank is looking for does
not exist in Somaliland. The model argues instead that the
redistribution of fiscal resources from the government to the
different regions lies at the heart of these peaceful and demo-
cratic institutions. The key role of redistribution in peaceful
state-building in Africa has been emphasized in particular by
Azam (2006). In Somaliland this redistribution of fiscal
resources has mainly funded a significant expansion in the
education sector in all regions of the country. Between 1997
and 2006 the number of primary schools rose from 165 to
516, while the number of universities increased from 1 to 5
(MNPC 2010).
The model presented later argues that the Somaliland
Republic shares many features of the “indirect rule” system
156 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND
that was widely used in the days of the British Empire and
that has been used in many developing countries since their
independence. In particular, Boone (2003) shows how this
system was applied quite successfully in several parts of
postcolonial Francophone West Africa, where the central
government delegated the task of controlling some areas to
some local traditional authorities, in return for some trans-
fers. I argue here that this system cuts through a vexing
“bootstrap” problem that faces all new states: a state needs to
have fiscal resources to extend its control to various parts of
its territory, but the state needs to have a fairly serious level
of control to be able to raise those fiscal resources in the first
place. This problem explains to a large extent why the con-
trol many African states exert is in fact extremely limited,
leaving de facto large parts of their country without any
effective state presence, as emphasized by Herbst (2000).
I argue that Somaliland was put on the fast track to solve
this problem thanks to two of its preexisting assets. First,
this country inherited a valuable transport infrastructure,
which only required establishing an efficient political
regime to become competitive in the Horn of Africa. I sug-
gest that “ports are the taxman’s best friends,” because they
provide a “choke point” where taxable resources are con-
centrated and make revenue collection relatively cheap. Sec-
ond, the traditional institutions of this nomadic pastoral
society had not been destroyed by either British colonial rule,
or the subsequent “modernizer” national government of
post-colonial Somalia, despite the brutal attacks by the
Mogadishu government in the final years of Siyad Barre’s
rule in the late 1980s (Lewis 2008). That indiscriminate vio-
lence against civilians and soldiers alike probably helped the
Somalilanders to achieve a consensus on the project to
secede from Somalia and build a state of their own. The clans
with which Somali nomadic herdsmen are affiliated are
themselves subdivided into kinship groups, which are sub-
ject quite informally but firmly to the leadership of the eld-
ers. There is no real “chiefdom” among the Somali, unlike in
many other African societies (Lewis 2008), but the elders
exert a significant level of authority.
This power was successfully harnessed to the emerging
Somaliland Republic by creating a House of the Elders,
called Gurti in Somali, in addition to a more standard
elected House of Representatives. This upper house is play-
ing a part in Somaliland’s bicameral system close to that
played by the House of Lords in Britain’s Westminster sys-
tem, allowing the traditional authorities to be involved
directly in running the country’s affairs. I argue that this is
one of the fundamental pillars of this country, because it is
the key to establishing—at a low up-front cost—the required
level of security for making the port of Berbera an economic
success and hence a reliable source of fiscal revenues. The
emerging Somalilander government is in fact delegating to
the elders the task of controlling violence and banditry, with
a view to protecting the traders who then pay taxes in return
for the transport services of Somaliland and its port. Then,
the redistribution of fiscal revenues pointed out by Eubank
(2010), as mentioned above, is the natural compensation
for the investment made by the elders in providing the key
public good that makes this lucrative trade possible. Hence,
my approach to Somaliland’s state-building shares some
features of the so-called “property rights” approach to the
theory of the firm (Hart 1995). The returns to the transport
infrastructure inherited by Somaliland thus depend cru-
cially on the “relationship-specific investments” (Hart 1995)
made by the elders in controlling violence and banditry,
which is in turn rewarded by some redistribution of fiscal
revenues. Similarly, Hart’s theory rests on “incomplete con-
tracts theory,” which assumes that only a fraction of the
observable information can be used as part of an enforce-
able contract, while the rest is not verifiable by a court,
although it is observable by the parties to the contract. My
model pushes this to the extreme, because there is no third
party that can be called upon to enforce any agreement
between the government and the clans’ elders.
The solution offered here to this fundamental commit-
ment problem brings out the key theoretical contribution
made by the current model relative to Alesina and Spo-
laore’s theory of the size of nations (Alesina and Spolaore
2003). These authors define the government as a country’s
monopoly producer of a public good that affects its people
differentially. They then raise serious doubts about the pos-
sibility of compensating people for these differential bene-
fits by transfers, because of the lack of commitment of the
democratic government that they assume. In the model
presented here, the government is unable to produce the
public good alone and must rely on the traditional author-
ities that are in a position to control violence and banditry,
provided their participation is satisfied through a transfer.
Then, the promise of this transfer can be made credible in
a repeated-game framework because the recipients can
punish any deviation by the government by reducing dras-
tically its payoff in case of cheating. This implicit threat is
credible because the recipient of the transfer is incurring a
positive opportunity cost in delivering its part of the deal.
Hence, what looks like a transfer is in fact the price paid for
a productive service sold to the government.
More generally, the state-formation theory sketched
below views the state as a means to internalize some key
CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND
externality that has the potential to enlarge the opportunity
set of the players, provided a fair compensation is paid to
the investors. This is why redistribution plays such a key
role in African state-building, as mentioned above. The rea-
son for this result is that the basic negative externality that
plagues African states is the threat of civil war or, more gen-
erally, the threat of violence. If I become armed, then your
expected welfare goes down, because there is a non-zero
probability that I will use these weapons to attack you. This
is what the “social contract” aims at preventing, by provid-
ing a fair and credible compensation for giving up one’s
weapons (Azam and Mesnard 2003). In Somaliland the
threat was more directed at the economy, because any inse-
curity felt by the traders would have brought the port of
Berbera to a halt. Nevertheless, this implicit threat was
overcome through a gradual bottom-up process leading to
the emergence of the democratic regime.
A MODEL OF TRADERS UNDER THREAT
After the breakup of Somalia, law and order collapsed, and
the country became prey to roving bandits and warlords,
making trade highly risky. However, Somaliland itself man-
aged to isolate a relative safe haven for traders. The following
model aims at bringing out the two levels of political organi-
zation that made this possible. I first analyze how the tradi-
tional system of social control was mobilized for reining in
uncontrolled violence. The next section shows how a higher
level of political cooperation was needed to create the
required level of security for making Berbera a success.
The traders
A very simple model was chosen for capturing the key part
played by security in determining the level of trade going
through Somaliland and Berbera. Let V be the value of the
goods transported through the country by the traders. The
traders potentially incur three types of costs while moving
across the country. First, there is a resource cost involved
in trucking the shipments, including fuel and labor. Sec-
ond, there is a possibility that bandits might rob the trader
along the way, leaving him without anything to sell at the
port; the probability that this happens is denoted as p.
Last, I assume that the lucky traders who have not been
raided have to pay a tax on the goods leaving the country
at the fixed rate t. On the export side, I thus assume that
the tax is paid at the port, and on the import side, at the
end of the trip, that is, mainly at the border for the transit
trade to Ethiopia. We assume that the government controls
157
 corruption well enough so that the tax rate is fixed before
other agents make their decisions and does not respond to
the observed trade flow coming through its control points.
All the parameters of the model are assumed to be common
knowledge, so that all the players can correctly anticipate
the decisions made later on by the other players. The
traders are assumed to play last, that is, to make their deci-
sion to start the trip or not while taking t and p as given.
It is realistic to assume that there is no free entry in the
trading business, mainly because of the limited warehouse
capacity and restricted credit that is typical in poor coun-
tries. In Africa long-distance traders usually belong to some
long-established family networks (Grégoire and Labazée
1993). Assuming for the sake of simplicity a quadratic cost
function, then the representative trader chooses Berbera
rather than any other port of the Red Sea or the Gulf of
Aden area if:
v2
E = max (1 − τ ) (1 − π ) v − ≥r (9.1)
v 2v
where r is the trader’s reservation profit, which he could
expect to earn while using an alternative trade route and v–
is the maximum carrying capacity of this route. From the
first-order condition, the amount of trade going through
the country if the weak inequality in (9.1) holds, is:
v = v (1 − τ ) (1 − π ) . (9.2)
Substituting back into (9.1), I find that a positive level of
trade will go through the country if:
1/2
(1 − τ ) (1 − π ) ≥ ⎛⎜ ⎞⎟ .
2r (9.3)
⎝v ⎠
Assuming that v– > 2r, (9.2) and (9.3) jointly imply that
when this is profitable, the level of trade will be such that:
( 2rv )
1/2
≤ v ≤ θ. (9.4)
Otherwise, it will fall to zero. This simple setting thus
captures the idea that the competitiveness of the port of
Berbera depends first on two parameters, namely, r and v–,
which determine respectively the profitability of doing busi-
ness with the competitors and the physical transport cost
inside the country, and then on two choice variables, p and t,
which are determined by two key players, respectively, the
bandits and the tax authority. In the following, I assume that
v– is much larger than 2r, in order to capture the geographi-
cal advantage of Berbera over its competitors. Then, Berbera
should win, provided the two key players manage to coordi-
nate their decisions efficiently.
158 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND
The potential bandits and the government
How would the potential bandits behave in a hypothetical
society where the traditional clan authorities would fail
somehow to organize their activity? I assume that there are
N bandits who sequentially choose whether or not to raid a
shipment. I make the simplest assumption regarding the
cost of raiding the traders, namely, that a fraction 0 < g < 1
of the shipment is lost in each raiding. This might capture
the collateral damage of any fighting between the bandits
and the traders, or any other form of cost that the bandits
incur in raiding. This assumption is the simplest, but most
results below are robust to several extensions. Given this
cost function, a given bandit i ∈{1, . . . ,N} takes the decision
xi ∈{0,1} to raid a trader to maximize his expected profit
simply defined as:
Bi = maxxi xip(1–g )vi, (9.5)
where p is the probability a raid is successful once under-
taken and vi is the value of the traffic faced by bandit i,
which takes value 0 if the trader has been successfully raided
earlier, and v otherwise.
This expression implies that the bandits incur no cost at
all if xivi = 0 either because there is no traffic to attack, or
because they have chosen not to undertake any raiding.
Otherwise, they loose a constant fraction of the catch when
they raid a trader. When the bandits are not organized, they
do not take into account the externality that they inflict on
other bandits when performing an attack, namely, that this
will in turn reduce the level of traffic that can be raided.
Thus, any bandit has an incentive to raid a trader, which
implies that the probability of a successful raid eventually
converges to 1 as the number of bandits grows large:
pN = 1–(1–p)N and lim π N = 1.
N →∞
Thus, the resulting expected value of the trader’s ship-
ment converges to zero:
lim v N = 0. (9.6)
N →∞
Against this background, I can now analyze the first con-
tribution made by the assembly of elders in the Somali clan
society in reducing banditry. In this nomadic society, where
the young men are spread all over the land in search of fod-
der for their flocks, collective identity is not defined so much
by reference to a territory as by genealogy. Moreover, in this
potentially violent society, where herdsmen could fight over
grass or water at any time without any witnesses, escalation
is prevented by the widespread use of “blood compensa-
tion,” or diya in the Somali language. This is a clever system
for creating joint liability within the group of origin of the
perpetrator of a violent crime, because the whole group is
responsible for paying the compensation required by the
victim’s group for settling the issue. According to Bradbury
(2008), the diya for the murder of one man is 100 camels
(half of this for one woman), which is a very high cost for
the group. If the diya is not paid, then the aggrieved group
is committed to launch a war against the criminal’s group.
This provides a strong incentive for those involved to exert
some control over their fellow clan members, so that any
man found guilty of a crime against a member of another
clan would potentially be punished by his own clan, thereby
avoiding interclan violence as far as possible. Hence, an
important service delivered by the clan is to control the vio-
lence that could be perpetrated by its own members against
both the clan’s members and the members of the other
clans. This gives the elders a key role in the Somali society’s
control of violence. This power was gradually aggregated in
Somaliland during the political buildup toward democracy
by organizing first a large series of local meetings of elders,
which then developed in a kind of pyramidal fashion, cul-
minating with the creation of the national assembly of the
elders at the Gurti. This process is well documented by
Bradbury (2008). The most striking point about this process
is that it was mainly organized and funded by the diaspora
of Somalilanders who had fled the repression under the
Siyad Barre regime. The diaspora played a key role in many
other parts of Somaliland’s political development and the
emergence of democracy there.
What can such a consolidation process deliver within my
model society of traders and bandits? Quite obviously, such
a consolidation process would end up creating a kind of
syndicated banditry, which could internalize the negative
externality involved because an increase in the raiding activ-
ity against traders would reduce the size of the trade flow
itself. Now, instead of problem (9.5), this coalition would
choose the number of raids or, equivalently given equation
(9.6), the probability of a successful raid so to maximize:
B S = max π (1 − γ ) v , (9.7)
π
such that (9.2) holds.
It can be readily checked that this “syndicated banditry”
equilibrium yields the following levels of raiding and traffic:
π S = 1 2 and v S = v (1 − τ ) 2 . (9.8)
This implies that the mere fact of forming a coalition of
clans is not sufficient in general to explain why raiding
would stop. Nevertheless, the model predicts that the cre-
ation of an institution that helps the traditional clan
CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND
authorities to coordinate their action makes a positive
contribution toward efficiency by reducing raiding. This
institution helps the clans’ leaders to internalize the nega-
tive externalities that they would inflict on one another by
raiding the traders without control, inducing them to
reduce their raiding activity.
Now, absent any political arrangement that could help
the government and the potential bandits to coordinate
their action, the government would simply maximize its
expected fiscal revenues, taking into account the traders’
best-response function (9.2) and the bandits’ best choice:
t ,x
( )
G = max τ 1 − π s v , s.t. (2) holds. (9.9)
Proposition 1 below describes the resulting Nash equilib-
rium that prevails in this model when the two players do not
coordinate their actions through some political arrangement.
PROPOSITION 1: The uncoordinated Nash equilibrium choice of
p and t by the government and the syndicated bandits,
respectively, is:
t N = p N = 1/2, (9.10)
entailing a level of traffic:
v N = v 4, (9.11)
and the following payoffs for the bandits and the government,
respectively, are:
G N = v 16 and B N = (1 − γ ) v 8 . (9.12)
The next section shows how a more inclusive political
arrangement can harness this social control mechanism
provided by the traditional clan authorities to improve effi-
ciency still further and reduce raiding to zero.
REDISTRIBUTION IN THE EFFICIENT
POLITICAL EQUILIBRIUM
In a clan society, genealogy is the essence of social identity,
as mentioned above. It is then natural to assume that the
elders have a strong interest in the continuation of the clan
and thus care for the welfare of the next generation. This
can be captured by using a dynastic family model à la Barro
(1974). In this kind of model, each generation is affected
by intergenerational altruism, such that the next genera-
tion’s welfare is an argument in the current generation’s util-
ity function. Choi and Bowles (2007) have coined the
expression “parochial altruism” to describe such an inter-
generational externality and have shown how these links
159
 across generations are an important asset for the survival of
human groups in a violent society within an evolutionary
framework. Hence, the dynastic family assumption seems
especially appropriate for describing the behavior of a tradi-
tional clan society like the Somali one. In this case, it is nat-
ural to assume that the players have an infinite horizon,
because their concern for the next generation creates a chain
of intergenerational links up to infinity. I thus discuss the
political setting in which banditry and warlordism can be
eradicated by embedding the simple model of the previous
section within an infinite-horizon repeated game frame-
work. I then show how Somaliland’s political institutions
cater for the key mechanisms brought out by this model.
To capture their common ethnic heritage in Somaliland,
assume that the potential bandits and the government have
the same discount factor 0 < d < 1. Then, assume that the
government can offer at each period the following contract
to the potential syndicated bandits: “I will give you g > 0 if
you refrain from raiding the traders and enforce p = 0”. This
contract clearly entails that the potential syndicated bandits
play first and the government second, after having observed
whether the raiding was avoided or not. Moreover, no third
party is available to enforce the promise made by the gov-
ernment to deliver the transfer once the potential bandits
have refrained from raiding the traders. I define an efficient
political equilibrium as an efficient outcome that can be
sustained ad infinitum in the repeated game between the
government and the potential bandits.
DEFINITION 1: An efficient political equilibrium is a triplet
{t, p, g} that lies on the Pareto frontier in the game
between the government and the potential syndicated
bandits and that can be sustained by a standard trigger-
strategy equilibrium.
In this simple setting involving a transfer, the Pareto-
efficient {t, p} pair of actions by the two players can be
derived by solving the following problem:
knowing that it will redistribute part of the resulting revenue,
as explained below, to compensate the potential bandits for
their restraint. Moreover, the chosen tax rate is the same as in
the Nash equilibrium of the previous section, at the level that
maximizes the Laffer curve t (1–t ). Then, the traffic level is
twice as large in this equilibrium as in the uncoordinated
Nash equilibrium of the previous section.
We can now prove the following:
PROPOSITION 2: There exists at least one efficient political equi-
librium with p ∗ = 0, t ∗ = 1/2 and:
4 − 3δ 3v δ
(1 − γ )v ⎛⎜ ⎞
⎟≤ g∗ ≤ , (9.15)
⎝ 8 ⎠ 16
if:
(1 − γ ) 8
δ ≥δ = . (9.16)
9 − 6γ
The proof is rejected in the appendix, while figure 9.1
helps the reader understand proposition 2 intuitively. The
downward sloping line represents the left-hand part of con-
dition (9.15). All the points located above this line are
acceptable for inducing the potential syndicated bandits to
cooperate. The required transfer is lower, the more patient
the potential bandits are. The upward sloping line repre-
sents the right-hand part of this condition; all the points
located below it are acceptable for inducing the government
to cooperate. The government is willing to pay more, the
more patient it is. Then, the figure makes it clear that the tri-
angle on the right labeled “acceptable triangle” is the set of
Figure 9.1 The Efficient Political Equilibrium Set
g
(1 –g )n

max ⎡⎣τ (1 − π ) + π (1 − γ ) ⎤⎦ v , s.t. (2) holds. (9.13) 2

t ,π

Using standard maximization techniques (Kuhn and

Tucker theorem), one finds easily that the Pareto-efficient 3n
outcome is a corner solution that implies the following: 16
Acceptable
π ∗ = 0 , τ ∗ = 1 2 and v ∗ = v 2 . (9.14) triangle

(1 –g )q
The intuition for this result is pretty straightforward. 8
Because of the unit cost g > 0, raiding is an inefficient way of d
d 1
collecting revenues from the traders. The government thus
will perform all the tax collection in the efficient equilibrium, Source: Author’s calculations.

160 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND

all the points that are acceptable to induce cooperation by
both players. Notice that d–, whose value is given at (9.16),
is strictly lower than 1. This means that the set of acceptable
{d, g∗} is not empty, ensuring existence of at least one pos-
sible efficient political equilibrium point. Moreover, d– is a
decreasing function of g, so that cooperation is easier to
achieve, the higher is the unit cost of raiding. Hence, (9.16)
defines a credibility frontier in the {d, g } space, which is
used in the next section. The intuition for this result is
that the stronger the comparative advantage of the govern-
ment at collecting revenues, the easier it is for it to buy off
credibly the potential bandits with a transfer for producing
security for the traders. The next section exploits this intu-
ition to diagnose why a solution similar to Somaliland’s
did not emerge elsewhere in Somalia.
WHY NOT THE REST OF SOMALIA?
The foregoing modeling exercise begs the question of why
such an efficient political equilibrium did not emerge in the
rest of Somalia. One easy answer is given by Eubank (2010),
who claims that foreign aid played a detrimental role there,
by relaxing the need to build accountable institutions in that
part of the country. Nevertheless, it is worth going into
more detail to look at the missed opportunities in eastern
and southern Somalia in order to bring out the kind of diag-
nosis that the model above is pointing out. It emphasizes
some more structural characteristics.
The first point to notice is that without declaring seces-
sion formally, the northeastern part of Somalia also built a
bottom-up institutional solution known as the “New
Puntland State of Somalia,” which was founded at a con-
ference in Garowe in 1998. This promising solution started
among some Darod clans, in particular the Mijerteyn
(Lewis 2008), but it was missing two of the key ingredients
of Somaliland’s success. First, that part of the country did
not inherit an infrastructure asset of the same caliber as
the port of Berbera. In fact, the Puntland ruling elite never
lost sight of the nearby formal capital city of Mogadishu in
the south. That is where the infrastructural assets are nat-
urally located, despite the massive destruction brought
about by the war. Mogadishu has two ports, one new and
the other old, and an international airport. Moreover a
road to Addis Ababa, called the “Strada Imperiale,” could
be restored. Hence, the Puntlanders never severed their
links to the rest of Somalia, realizing probably that they
would never be in a position to levy the fiscal resources
required to cement a Puntlander social contract similar to
the one prevailing in Somaliland. Their strategy was clearly
CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND
leaning in the opposite direction, suggesting that they were
just regarding the “New Puntland State” as a mere building
block to reconstruct Somalia. Second, they did not build any
institutional representation of the elders, because they cre-
ated a unicameral parliament, which did not have many
resources in any event. In terms of the model described
above, this lack of resources might be blamed on the lack of
significant transport infrastructures and thus on the lower
need for internal security. This is illustrated by the fact that
piracy is quite active off the coast of Puntland, imposing
some negative externality on all the ports of the Red Sea
and the Gulf of Aden. However, the failure to give the eld-
ers a large enough role in the New State of Puntland is also
probably due to the towering figure of Colonel Abdillahi
Yusuf, who assumed first the presidency in Puntland, up to
2001, and then won the federal presidency in Mogadishu,
which is in fact an empty shell. Abdillahi obviously did
not draw the lessons of the failure of Siyad Barre’s military
regime to create top-down a viable state in Somalia, because
he tried also to impose his authoritarian rule on the New
State of Puntland. However, imposing the “rule of fear” also
requires resources, which are dramatically lacking in
Puntland, while they are somewhat higher in Mogadishu,
if only because of foreign aid as mentioned above.
Southern Somalia once had the resources to support
the strong authoritarian government led by Siyad Barre,
who for a while was aligned with the Soviet Union, which
was playing a complicated game in the Horn of Africa.
Then alliances switched, and the United States became
involved. Although southern Somalia inherited a valuable
infrastructural asset, in addition to the sovereignty rent due
to international recognition and foreign aid, it was facing a
more complicated political problem than Somaliland.
While the latter is very homogenous, with most of its pop-
ulation involved in nomadic pastoral activity, the former
has a sizable agricultural area, between the Shabelle and
Jubba rivers. The Somali clans living there have a distinct
sedentary culture, in which territory matters at least as
much as genealogy, so the elders have a weakened role in
social control. Moreover, there are some Bantu farmers in
the midst of this Somali population, loosening further the
ethnic ties in that part of Somalia. Hence, the traditional
authorities are too weak in that part of Somalia to deliver
the kind of social-control services available in Somaliland,
on the one hand, while the government was too authori-
tarian to make credible promises of redistribution, for lack
of checks and balances, on the other hand.
Figure 9.2 summarizes the foregoing discussion within
the analytical framework presented earlier. It represents
161
 Figure 9.2 Why Not the Rest of Somalia
d
1 CONCLUSION
8/9 Somaliland
Other
Somalia
g solution in some circumstances, for “breaking up” a state of
1
Source: Author’s depiction.
condition (9.16) as the downward sloping frontier repre-
sented in the {d, g } space. This is the credibility frontier
defined above. The parameter g measures the cost advantage
enjoyed by the Somalilander government over the potential
bandits in collecting revenue, thanks to the preserved stock
of infrastructure inherited from colonial days, and enhanced
over the recent past by some investment. In the case of
Puntland, no such stock is available, as explained above,
suggesting that g is small there, because the government is
not in a position to collect revenue much more efficiently
than the potential bandits. In Southern Somalia, g is poten-
tially much larger, because of the infrastructure available in
Mogadishu, although the latter needs massive investments to
recover its potential efficiency. However, the mixed ethnic
composition of the population living between the Jubba and
the Shabelle rivers, as well as the sedentary culture of the
Somali agriculturists living there, suggest that d is much
lower there than in Somaliland. Hence, this analysis empha-
sizes that Somaliland exploited the two-dimensional edge
that it had over the rest of Somalia for creating its efficient
political equilibrium. This is captured in figure 9.2 by noting
that Somaliland’s efficient political equilibrium prevails for a
{d, g } pair lying above the frontier, while the other parts of
Somalia are found below that frontier.
This finding also suggests that simply trying to export
the solution that worked in Somaliland to the rest of Soma-
lia would probably not bring about the same benefits. Some
additional imagination is surely needed to devise an appro-
priate system for the rest of Somalia with a view to create a
lasting peace there. Nevertheless, the analysis presented
here points to the two dimensions where economic and
political innovation is required: enhancing the relative effi-
ciency of the government at collecting revenue, relative to
warlords and bandits; and making credible the necessary
162 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND
redistribution mechanisms for compensating the latter for
giving up their highly lucrative violent activities.
This case study of the emerging Somaliland Republic pro-
vides a natural experiment that sheds some useful light on
the theory of state-building. It shows that the Hobbesian
Leviathan is not the only path available for controlling vio-
lence and building up a peaceful state. It suggests that a
Lockean “horizontal social contract” model may be a viable
anarchy, using the expression coined by Hirshleifer (1995).
In Somaliland, one observes a separation of the power to
control violence, which belongs to the clans’ elders, on the
one hand, and the power to tax and to produce some of the
public goods that a modern state is expected to provide, on
the other hand. Among other things, this study thus
shows the benefit that political economists can gain from
using the work of the social anthropologists for under-
standing the political economy of developing countries. In
return, the modeling exercise demonstrates the key part
played by an inherited infrastructural asset, namely the
port of Berbera and the road that links it to Ethiopia.
This model sheds some light on the political institutions
that have been put in place in Somaliland. The key problem
to be solved was for the business-oriented elite to delegate to
the traditional authorities the task of controlling violence
and banditry effectively so that Berbera became an attrac-
tive outlet for the traders from Ethiopia as well as for
exporting the output of the livestock sector. The first step
was to help the clans’ elders to cooperate by organizing sev-
eral local conferences. But the model shows that this is not
enough to provide the incentives for reducing banditry to
zero. The second step was aimed at making credible the
promise of redistributing the enhanced fiscal resources
resulting from the increased trade flow to the clans. A
bicameral system was put in place to ensure that the elders
had a key role to play in the law-making process, by giving
them a direct access to the required information, as well as
some veto power in the implicit bargaining problem. As
noted, the redistribution of the fiscal resources was focused
on the development of the education sector in all the
regions, as one would expect in a country where genealogy
is the key social identifier that determines each person’s
affiliation to a clan. The dynastic family model thus seems
to apply perfectly in this case and explains why people felt
compensated for their efforts by seeing their children going
to school and to university.
 This model may thus be viewed as an extension of Hart’s
“Property Rights Approach” to the theory of the firm (Hart
1995). The clan authorities can invest in providing security,
an asset that enters the production function for transport
services as a complement to infrastructure. This insight is
fundamental for understanding why a bicameral demo-
cratic institution lies at the heart of Somaliland’s political
institutions, for providing a balanced representation of both
the traditional authorities and the business-oriented mod-
ern actors. Hence, Somaliland’s experience provides a fruit-
ful line of arguments in favor of a qualified support for the
traditional “project aid” doctrine, with its emphasis on
funding infrastructural projects, which has inspired the
action of the World Bank and other development agencies
CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND
for decades. The qualification brought out by this case study
is that a correct political setting is required, aimed at mak-
ing the redistribution of the benefits from cooperation
among the different actors credible. This redistribution is
the compensation due to the potential bandits for refraining
from raiding the traders and thus participating in the effi-
cient political equilibrium. This suggests that the inability to
set up a correct political system, partly because of external
interference, is what makes the infrastructural assets of
southern Somalia, like Mogadishu’s old and new ports, its
airport, and the “Strada Imperiale” road that links them to
Addis Ababa, as well as the rent to sovereignty provided by
access to foreign aid that comes with international recogni-
tion, largely useless for producing peace.
163
 APPENDIX PROOF OF PROPOSITION 2 2. Markakis (1992) provides a regional perspective on the
ports in the Horn of Africa.
Assume that both players adopt the standard trigger strategy
(see, for example, Gibbons 1992). If they choose to cooper-
ate and refrain from raiding the traders with a view to REFERENCES
receiving the transfer g from the government, the potential
Alesina, Alberto, and Enrico Spolaore. 2003. The Size of
syndicated bandits receive the following present value:
Nations. Cambridge, MA: MIT Press.
g
VC = . (A9.1) Azam, Jean-Paul. 2006. “The Paradox of Power Reconsid-
1−δ
ered: A Theory of Political Regimes in Africa.” Journal of
If they choose instead to deviate and attack the traders, African Economies 15 (1): 26–58.
the syndicated bandits will be punished first within the Azam, Jean-Paul, and Alice Mesnard. 2003. “Civil War and
same period, as the government withholds the transfer g, the Social Contract.” Public Choice 115 (3): 455–75.
and the static Nash equilibrium outcome will then prevail Barro, R. J. 1994. “Are Government Bonds Net Wealth?”
ever after. Assuming that this deviation is not expected by Journal of Political Economy 81: 1095–1117.
the traders, this yields the following present value: Boone, Catherine. 2003. Political Topographies of the African
(1 − γ ) v δ (1 − γ ) v State: Territorial Authority and Institutional Choice. Cam-
VD = + . (A9.2) bridge, U.K.: Cambridge University Press.
2 8 (1 − δ )
Bradbury, Mark. 2008. Becoming Somaliland. London:
The potential syndicated bandits will thus choose to Progressio.
refrain from raiding the traders if VC ≥ VD, that is, if: Choi, Jung-Kyoo, and Samuel Bowles. Samuel. 2007. “The
⎛ 4 − 3δ ⎞ Coevolution of Parochial Altruism and War.” Science 318
g ≥ (1 − γ ) v ⎜ ⎟. (A9.3) no. 5850: 636–40.
⎝ 8 ⎠
Doornbos, Martin. 1993. “Pasture and Polis: The Roots of
On its part, if the government chooses to cooperate, that Political Marginalization of Somali Pastoralism.” In
is, to deliver the agreed amount g when observing p = 0, it Conflict and the Decline of Pastoralism in the Horn of
receives the following present value: Africa, ed. John Markakis, 100–121. Basingstoke, U.K.:
v − 4g Macmillan.
WC = . (A9.4)
4 (1 − δ ) Eubank, Nicholas. 2010. “Peace-Building without Exter-
nal Assistance: Lessons from Somaliland.” Working
If it chooses instead to deviate and to withdraw its trans- Paper 198, Center for Global Development, Washing-
fer despite the potential bandits’ compliance with the ton, DC.
promised p = 0, then the government will first keep the Gibbons, Robert. 1992. Game Theory for Applied Economists.
whole fiscal revenue in the current period, and then get its Princeton NJ: Princeton University Press.
Nash-equilibrium payoff GN from then on. This yields the Grégoire, Emmanuel, and Pascal Labazée, eds. 1993. Grands
following present value: Commerçants d’Afrique de l’Ouest. Paris: Karthala-Orstom.
v δv Hart, Oliver. 1995. Firms, Contracts, and Financial Structure.
WD = + . (A9.5)
4 16 (1 − δ ) Oxford, U.K.: Oxford University Press.
Herbst, Jeffrey. 2000. States and Power in Africa. Compara-
Then, the government will choose to cooperate if
tive Lessons in Authority and Control. Princeton, NJ:
WC ≥ WD, that is, if:
Princeton University Press.
3θδ
g≤ . (A9.6) Hirshleifer, Jack. 1995. “Anarchy and its Breakdown.” Journal
16 of Political Economy 103 (1): 26–52.
Then (A9.3) and (A9.6) together make (9.16). Last, d_ is Lewis, Ioan M. 2008. Understanding Somalia and Somaliland
computed as the value of d such that the range of values of New York: Columbia University Press.
g∗ defined at (9.15) is just empty, as illustrated in figure 9.1. Markakis, John. 1992. “The Regional Significance of Sea
Ports in the Horn of Africa.” In Beyond Conflict in the
Horn, eds. Martin Doornbos, Abdel Gaffar M. Ahmed,
NOTES
and John Markakis, 130–31. London: James Currey.
1. “Bolloré proposes Berbera-Addis Ababa Transport Cor- MNPC (Ministry of National Planning and Coordination).
ridor.” WorldCargo News online. December 10, 2009. 2010. Somaliland in Figures 2009. Hargeysa.

164 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND

PA R T I I I

Leveraging Sectoral Advantages

to Expand Exports
 CHAPTER 10
Growing Mali’s Mango Exports: Linking
Farmers to Market through Innovations
in the Value Chain
Yéyandé Sangho, Patrick Labaste, and Christophe Ravry
lthough less than 4 percent of land in Mali, a
A
landlocked country in West Africa, is arable, agri-
culture accounts for 45 percent of the country’s
economy and employs 80 percent of its workforce. Indus-
try represents 17 percent of the country’s gross domestic
product (GDP), with food processing, construction, and
phosphate and gold mining as the principal industrial
activities. Mali’s main agricultural export is cotton, followed
by livestock. The fact that Mali is landlocked, however, has
always made it dependent on the transport infrastructure
and other logistical arrangements of its neighbors for trade
and exports.
Thanks to excellent geographical and weather conditions
prevailing in the southern part of the country, mangoes
have always been abundant in Mali, particularly in the
Bamako and Sikasso regions. The fruit was traditionally sold
in the domestic market. In the 1970s Mali was the first
country in West Africa to explore opportunities to export
fresh mangoes. These exports were shipped exclusively by
air freight to a niche market—high-end retail shops in
France selling tropical fruits—and reached a volume of
between 1,000 and 1,500 metric tons a year
Starting in the early 1990s Mali undertook several
transformations in its mango subsector that have allowed
the country to overcome logistical dependencies and con-
straints, expand exports of fresh mangoes, and make
Yéyandé Sangho is Senior Operations Officer at the World Bank. Patrick Labaste is Lead Agriculture Economist at the World Bank. Christophe Ravry is
Senior Agribusiness Specialist at the World Bank. This paper builds on an early draft and research work done by Malick Antoine at the World Bank in 2006.
major leaps toward developing a competitive horticulture
export sector.
MALI’S EXPORT DIVERSIFICATION STRATEGY
A key objective of Mali’s poverty reduction strategy over the
past two decades has been—and still is—to increase rural
incomes and employment opportunities by promoting agri-
cultural diversification and developing exports of high-
value commodities. In the early 1990s the government of
Mali recognized a need to design policies to diversify
exports and foreign exchange earnings, which had been
heavily dependent on only three export products: gold, cot-
ton, and livestock. These products, however, are quite sus-
ceptible to fluctuations. For example, cotton exports
dropped dramatically in 2008 to less than half of their pre-
vious level; also, because of several years of financial crisis,
the contribution of the sector to fiscal revenues has been
negative throughout the whole decade.
Striving for diversification
In the face of these fluctuations, the government of Mali
began, in the 1990s, to focus on high-value and nontradi-
tional agricultural products as a means to generate income
and achieve greater diversification of exports, based on the
167
 country’s comparative advantages. Besides offering small-
holder farmers the opportunity to diversify the source of
their livelihoods, high-value and nontraditional agricultural
products also offer countries the opportunity to diversify
away from low-value bulk commodities.
Several horticulture crops were considered as possible
targets for these diversification efforts, including cashews,
tomatoes, shallots, and mangoes. Mangoes were a prime
candidate both because of the excellent agroclimatic condi-
tions for growing them in the southern regions of Bougouni
and Sikasso and because of the fast-growing demand for
mangoes in European markets (figure 10.1). Furthermore,
because mangoes were already being produced by small-
holder farmers throughout the country, the subsector had
the potential to contribute to rural livelihood improve-
ments. However, despite the high quality of Mali’s fresh fruit
and vegetables, the high cost of air freight was severely lim-
iting marketing and exportation. In fact, significant volumes
of Mali’s mangoes were purchased and processed for export
by operators based in Côte d’Ivoire, thus leaving little
potential for value addition in Mali.
Early efforts toward agricultural diversification
In 1992 Mali’s Ministry of Agriculture prepared a national
rural development strategy, the Schéma Directeur du
Développement Rural (SDDR), emphasizing commercial
agriculture, export promotion, and value addition, and
the government began directing resources toward those
ends using financing from international donors such as
the World Bank and the U.S. Agency for International
Development (USAID).
Figure 10.1 Imports of Mangoes into the European
Union, 2004–08
250,000
200,000
Metric tons
150,000
100,000
50,000
0 Within Africa, leading producers, in order, are South
2004 2005 2006 2007 2008
Source: Eurostat.
168 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
In 1996 the government of Mali used $6 million in World
Bank financing for an agricultural trading and processing
pilot project, PAVCOPA (Project d’appui à la valorisation et
à la commercialisation des produits agricoles). The project
sought to promote agribusiness and exports. Specifically, it
aimed to “improve the enabling environment needed to
enhance private sector business opportunities and encour-
age [state] disengagement from commercial activities,
improve technical support to producers through effective
research and extension aimed at enhancing and diversifying
production and improving the international competitive-
ness of Malian exports, and boost private investment in
agricultural trading and processing” (World Bank 1996).
In addition, the PAVCOPA project aimed at supporting
producers, processors, and traders in the Sikasso, Segou,
and Koulikoro regions and the Bamako district, providing
technical and promotional assistance (including price
information) and organizing commercial forums, sup-
porting professional associations, and carrying out studies
on the markets for high-value crops.
The Agricultural Value Chain Promotion Agency
(APROFA, by its French acronym), established by the gov-
ernment in 1993, was designated as the project executing
agency for PAVCOPA. APROFA’s goal was to create sustain-
able growth in the agrifood sector by increasing exports to
the European and African regional markets and establishing
import substitution activities. Further, the agency was to
support the improvement of the technical, managerial,
organizational, and professional capacities of public and
private actors engaged in agribusiness.
The global market for fresh mangoes
Demand for mangoes in the European Union (EU) has
grown significantly in recent years, increasing by approxi-
mately 55 percent between 2001 and 2008, from 136,000
tons to more than 230,000 tons respectively (table 10.1).
The leading exporter countries, in order, are Mexico, Brazil,
Peru, India, Pakistan, and the Philippines; together, they
represent around 75 percent of the market. Although
African mango exports grew by 69 percent during the same
period, export volumes remain far lower than those of the
leading producers. In 2007, for example, Mexico was
exporting a total of 236,000 metric tons of mangoes, while
the entire African continent recorded 46,300 metrics tons of
exports only (FAOSTAT).
Africa, Côte d’Ivoire, Sudan, Kenya, the Arab Republic of
Egypt, and Mali. Historically, Côte d’Ivoire has exported
 Table 10.1 Mango Imports to the European Union, by Country, 2004–09
(metric tons)
Growth 2004–08
(percent)
Country 2004 2005 2006 2007 2008 2009a Annual Total
Brazil 69,319 82,293 84,858 82,993 96,870 69,590 9 40
Peru 19,817 26,394 41,027 36,854 50,756 25,062 27 156
Pakistan 10,938 12,306 10,120 13,224 12,941 12,913 4 18
Israel 8,059 12,548 11,181 14,808 12,261 12,606 11 52
Côte d’Ivoire 11,426 9,856 14,428 14,706 11,249 11,659 0 –2
United States 7,612 6,894 5,971 7,404 7,516 5,536 0 –1
Senegal 2,810 3,011 6,194 4,702 6,034 6,219 21 115
Costa Rica 3,983 6,271 7,545 4,664 5,360 5,685 8 35
Mali 2,096 2,560 3,477 4,317 4,902 3,480 24 134
Dominican Rep. 1,228 1,591 1,618 2,767 4,307 4,179 37 251
India 915 1,720 2,472 2,425 2,577 2,470 30 182
Burkina Faso 928 1,164 2,152 3,191 2,406 1,957 27 159
Other 23,516 20,035 19,786 19,003 13,209 12,893 –13 –44
Total 162,646 186,643 210,829 211,057 230,388 174,248 9 42

Source: Eurostat.
a. January to November.

significantly higher volumes of mangoes than its closest
competitors on the continent, apart from South Africa. In
the early 2000s, before political disturbances in the country,
Côte d’Ivoire exported around 11,000 metric tons; the next
closest African rival was Burkina Faso with 3,500 metric
tons exported. That same year Mali reported 900 metric
tons of exports (FAOSTAT).
THE CHALLENGES
Identifying market opportunities
The large amount of mango exports from Côte d’Ivoire can
be partly explained by the fact that the country was working
with buying agents from Mali in the 1980s and 1990s.
Mango exporters in Côte d’Ivoire operated pack houses in
the northern part of the country, sourcing mangoes from
growers across the border in Mali. The arrangement led to a
surge of mango exports, sold under an Ivorian label.
Malian mango producers did not anticipate the increas-
ing demand for mangoes in the European market, which
was moving from a luxury, niche market to a volume
market. Mangoes, like some other tropical fruits, such as
bananas and pineapples some decades earlier, became a
fruit that was in high demand all year round by European
customers.
In an attempt to increase its mango exports, Mali sought
alternatives to air freight that would increase its competi-
tiveness and market share. Mali’s landlocked status posed a
serious challenge to that effort, however, and logistics and
transport issues prevented an initial substantive scaling up
of exports.
Overcoming transport and logistics constraints
High transportation costs made Malian mangoes uncom-
petitive in the global market. Whereas competitors in Latin
America could take advantage of more economic sea freight,
Mali’s producers were limited to the more expensive air
freight option, which reduced their exports and relative
position as a mango exporter in West Africa (table 10.2).
Historically, Mali relied on Côte d’Ivoire as a link to port
facilities. Cotton lint and other products, for example, were
exported through the port of Abidjan. In addition to relying
on its neighbor’s ports, Mali also relied on its neighbor’s
infrastructure to move its export products to the ports.
Until the 1990s the only rail line in Mali with international
links was run inefficiently, leading to significant high prices
and delays. This protracted transport process was especially
detrimental to those Malian agricultural products that have
high spoilage and shrinkage rates. A research report
(CARANA Corporation 2004) also points out that produc-
ers’ logistical costs were negatively affected by the cost of the
consolidation of goods and poor business practices.
Other challenges
Although infrastructure was the most serious constraint to
Malian exports of fresh produce, inadequate access to
finance and land; a poor business climate; inadequate

CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 169

Table 10.2 Mango Exports to the European
Union, 1970–95
(metric tons)
Mali Burkina Faso Côte d’Ivoire
1970 35 122 —
1980 1,172 2,116 281
1990 1,300 2,700 1,000
1995 850 714 7,107
Source: FAOSTAT.
Note: — = not available.
management, harvesting, and handling techniques; and
little investment in mass production further weakened
producer export capacity.
Lack of Market Information and Organization.
Malian producers and exporters did not have a good grasp
of the requirements of the intensely competitive European
market. In addition, poor organization and coordination
between producers, government inspection officials, and
exporters led to inefficiencies affecting the industry’s ability
to meet international orders on time.
Lack of Proper Harvesting Practices and Posthar-
vest Handling Techniques. Buying agents typically vis-
ited orchards once a season and harvested all of the fruit at
once, regardless of its ripeness. In addition, many growers
did not manage their orchards, did not prune undergrowth,
and did not clean the ground under the trees, creating a
good habitat for fruit flies.
Little Investment at the Production Level. Invest-
ment in mango production—including the establishment of
commercial orchards—is important because it provides
producers and exporters a degree of organization and effi-
ciency. Traditionally, Malian farmers saw mango production
as mainly a subsistence activity, a business opportunity.
Even though orchards were smaller than five hectares, as
much as 50 percent of total mango production was wasted
each year (Club du Sahel/OECD 1998), and little investment
was made in increasing production. Additionally, because
Malian mangoes were exported through Ivoirian pack
houses, Malian producers received little return from their
crop and thus had even fewer incentives to invest in com-
mercializing crop production.
An Uninviting Investment Climate. Foreign investors,
unfamiliar with Mali’s policies and procedures for conducting
business, perceived investment in the country as high risk. In
170 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
addition, limited foreign direct investment in the country and
poor enforcement of fair business practices further reduced
competitiveness.
Poor Land Titling and a Nonexistent Land Market.
Bureaucratic inefficiencies and land tenure rules kept
Malian producers from owning large tracts of land, thereby
creating another barrier to large-scale horticulture. Until
the Agriculture Orientation Law of 2006, there were no pro-
visions for the establishment of commercial-scale irrigated
plots, nor could producers purchase irrigated land that
could be used as collateral against which to secure loans.
Inadequate Working Capital and Lack of Supply
Chain Financing. With limited working capital, Malian
production and subsequent exports were constrained because
exporters were required to finance the cost of shipping. This
problem was of particular urgency to exporters, who did
not always receive payments from previous shipments before
they needed to send additional shipments.
In short, not only did Mali need to make changes that would
enable it to improve its main constraint—infrastructure—it
also needed to undertake a series of transformations improv-
ing technical capacity and business regulatory practices in
the mango sector (box 10.1).
INTERVENTIONS
Supporting agricultural diversification
The transformation of Mali’s mango export sector began in
the mid-1990s with PAVCOPA, the pilot project funded by
Box 10.1 The Mango Export Value Chain before
Reforms
Before reforms in the sector, the value chain included
producers, intermediaries, and exporters.
■ Producers sold produce to buying agents, called
“pisteurs,” independent traders who select, harvest,
wash, sort, package, and transport the fruit from
orchard to pack house.
■ Agents worked exclusively for a particular exporter.
■ Quality control was undertaken by exporters, who
selected fruit that were of export quality and
returned the second-grade fruit to the pisteur for
sale on the domestic market.
Source: Authors.
the government of Mali and the World Bank. Initially,
Malian buying agents used existing facilities at the Bamako
airport, to export about 1,500 metric tons of mangoes to the
European market each year. In the early 1990s, however, as
demand in Europe increased, competition intensified
among producers as South American exporters, notably
Brazil, began exporting sea-freighted mangoes to Europe on
a large scale. Even though air freighting allows mangoes to
be picked at a more advanced stage of maturity (resulting in
sweeter fruit), it also comes at a significantly higher cost.
In addition to the government and World Bank initia-
tives, USAID contributed to jump-starting initial work in
Mali’s mango subsector through support to the Office de la
Haute Vallée du Niger (OHVN) project, that also focused on
high-value agriculture and food products (HVAFs). This
program served as a pilot project while the preparation of
PAVCOPA was ongoing. Aid organizations continued to
be involved in Mali’s agriculture sector for over 15 years
following the initial project in 1993.
Rethinking and redesigning the
mango export value chain
The first step that triggered the agricultural export reform
process in Mali occurred when a small but critical mass of
stakeholders were brought together to thoroughly analyze
and assess opportunities for diversification into higher-
value crops. Initial implementation of PAVCOPA was not
unsuccessful; however, the preparation of its midterm
review in January 2000 provided an opportunity to restruc-
ture the project and create a new business plan.
Another critical point was reached when APROFA and
its technical team of advisors began analyzing what had
made Côte d’Ivoire successful in developing sea-freighted
mango exports. Although Mali had been exporting mangoes
for years, the industry was not broadly based but rather lim-
ited to a narrow market targeting an exclusive club of
exporters on the periphery of Bamako.
Observing and understanding the success of Côte
d’Ivoire entailed a thorough analysis of the global market
demand and its trends, and an assessment of the supply
chain in Côte d’Ivoire. This evaluation suggested that three
critical steps were necessary to change the mango industry
in Mali. The industry had to penetrate and compete in the
growing market for fresh mangoes (that is, the sea-freighted
trade to major European ports and hubs such as Rotterdam,
Antwerp, Algesiras, and London); establish the drivers of
change and a value chain within the system to improve
competitiveness; and identify and implement solutions to
CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
overcome poor access to international ports while develop-
ing alternate and effective transport routes.
A key innovation: Introduction of multimodal
transport in the export value chain
The design and implementation of a multimodal trans-
portation system in 1995 was a key innovation for the Malian
mango industry that eased transportation logistics con-
straints and costs. Through this new system, mangoes are
loaded into temperature-controlled containers and driven
into Ferkessedougou, Côte d’Ivoire. There, the containers are
transported by rail to Abidjan and shipped to Europe. The
containers are maintained at 5 degrees Celsius. Transit times
per shipment between the mango-producing regions of
Sikasso and northern Europe have declined from 25–30 days
to approximately 12–15 days (Danielou, Labaste, and Voisard
2003). As figure 10.2 illustrates, multimodal transit also
allowed large volumes of mangoes to be exported from Mali
at an affordable cost. The existence of this uninterrupted
cold chain (refrigerated containers) greatly decreases the rate
of spoiling of mangoes during transit.
Designing the innovation and testing the system
To hedge risks involved with the new system, APROFA nego-
tiated and signed a partnership agreement with a privately
owned Ivorian company, SN Tropical Expressions (SNTE),
that split the costs, profit margins, and risks associated with
the first shipment of 200 tons. SNTE was responsible for
both the logistics and packaging involved in shipping the
produce to their ultimate port of destination in Europe. A
profit margin was built into the fixed price to ensure that
both parties respect their mutual costs objectives. If the aver-
age selling price is greater than the fixed price, APROFA and
SNTE share the difference. This partnership resolved the
problem of lack of finance for exporters because SNTE had
the necessary working capital to fund the operation.
Under the same agreement, APROFA was able to obtain
a guarantee from a local bank enabling SNTE to lease an
existing pack house in Sikasso with the necessary equip-
ment for precooling fruit pallets. In addition, SNTE sec-
onded two experienced managers to operate the Sikasso
facility. Precooling the mangoes at the pack house stage is
critical for the continuous cold chain. The establishment of
continuous cold chains in the export of horticulture and
other high-value crops remains one the most serious chal-
lenges to the expansion of exports from other Sub-Saharan
African countries.
171
 Figure 10.2 Supply Chains for Fresh Mango Exports from Mali

a. Air b. Truck/sea
Air Truck
Farm Smallholder mango producers Farm Smallholder mango producers
M M
a a Small traders
Small traders l Traders
l Traders (pisteurs)
(pisteurs) i
i
Pack houses /
Packers/exporters Pack houses / Packers/exporters exporters
exporters

D
E C i
u ô v
r o Exporters
t i
o e r
Importers Exporters
p e
e E
2 days u
High cost, r Importers 25 days
low volume o
p Lowest cost,
e Wholesalers Retailers low volume

c. Multimodal
Multi-Modal
Smallholder mango producers
Farm

M Small traders
a Traders (pisteurs)
l
i
Pack houses /
Packers/exporters exporters

D Ferke
C i train station
ô v
t o Exporters
i
e r
e Exporters

E
u
r Importers 12 days
o Low cost,
p high volume
Wholesalers Retailers
e

Source: Geomar International.

Early success quality: none of the 63,000 cartons of mangoes shipped to

the Netherlands was rejected. To the contrary, there were
The test carried out by APROFA and SNTE was a great
demands for steady shipments of Malian mangoes. The
success from many points of view, particularly in terms of

172 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
validation of the feasibility and profitability of the new
value chain for mango exports represented a critical break-
through and opened completely new horizons for the
mango business in Mali.
Despite sociopolitical turmoil in Côte d’Ivoire, Mali’s
cold chain transportation system remained uninterrupted.
The pilot project helped Malian producers and exporters
address the three major constraints of Mali’s export supply
chain: finance, management, and infrastructure. Producers
were able to receive a higher price for their mangoes at the
farm gate—125 CFA francs versus 50 before, a 150 percent
increase. Similarly, exporters increased the volume of fruits
they were able to ship from Mali, which translated into
increased revenues. The impact of the pilot project has been
far reaching and has encouraged entrepreneurs to engage in
similar ventures within and outside the mango sector.
Although the full export potential of Malian mangoes is
yet to be determined, APROFA conducted a number of
studies to produce estimates of such figures. In the Sikasso
region, for instance, where production in 2001 was esti-
mated at 48,000 metric tons, it is expected that the region
could export 14,400 metric tons (under the assumption that
only 30 percent of the mangoes would be of export quality).
APROFA estimates that with total production of approxi-
mately 200,000 metric tons, Mali would be able to export
50,000 metric tons (25 percent) of its mangoes.
Consolidating success and transforming
agricultural exports in Mali
In 2005 the government of Mali launched the Agricultural
Competitiveness and Diversification Project, a six-year
investment project, funded by a $46.4 million World Bank
credit. The project’s goal is to increase revenues and com-
petitiveness for a range of agricultural products with grow-
ing (yet mostly untapped) markets and strong demand, thus
diversifying the economy’s foreign exchange earnings. In
addition, its development objective is to improve the per-
formance of supply chains across a range of nontraditional
agricultural, livestock, fisheries, and produce in which Mali
enjoys a strong competitive advantage, including mangoes,
cashews, shallots, potatoes, dairy products, beans, papayas,
sesame, and shea nuts.
The project includes five components: demonstration
and dissemination of irrigation, postharvest, and value-
adding technologies; improvement of the performance of
existing and developing supply chains; facilitation of
access to finance for producers and operators; investment
in key collective, market-oriented infrastructure; and
CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
management, monitoring, and evaluation of project
implementation.
Using a series of analytical tools and data, the project
team identified value chains for export markets, providing
a basis for prioritized interventions that would ultimately
create value and improve livelihoods of Malian small farm-
ers. The analysis used five modules to assess the competi-
tiveness of a range of nontraditional agricultural value
chains (box 10.2). Each module built on the previous one
and progressed from a comprehensive list of sectors to
those with true marketability, competitive advantage, and
comparative advantage. This process also took into account
the demand in existing end markets and identified new
potential end markets, regional climate and growing factors,
production capacity, access to finance, and infrastructure,
among other determinants.
Improving quality by investing in
infrastructure and technical assistance
Mali’s revised surface transport system vastly increased the
country’s export capacity for mangoes. In fact, once the new
transportation system was established and proven economi-
cally efficient and reliable, the government pursued additional
Box 10.2 The Strategic Profile Approach
The approach that was followed to identify and prior-
itize value chain interventions in Mali entailed the fol-
lowing sequence of modules, each one comprising
several steps:
■ Module 1: Defining Mali’s Broad Portfolio of Agri-
cultural Sectors
■ Module 2: Analyzing Market Demand and Market
Entry Conditions
■ Module 3: Analyzing the Competitiveness of
Potential Malian Offerings
■ Module 4: Defining Priority Sectors
■ Module 5: Competitiveness Planning: Putting the
Analysis into Action
Through this process, a series of operational tools
was developed for each of the selected value chains—
the Strategic Development Plan, the Competitiveness
Plan, and the Priority Action Plan. PCDA is now in
the process of implementing these action plans with
the respective value chain stakeholders.
Source: Authors.
173
 interventions to increase the quality and quantity of man-
goes exported. A number of successive agricultural diversi-
fication projects were thus started during the 2000s:
USAID’s Centre Agro-Entreprise (CAE), Trade-Mali, the
Initiatives Intégrées pour la Croissance Economique du
Mali (IICEM) program, and the World Bank’s Programme
Compétitivité et Diversification Agricoles (PCDA). Inter-
ventions covered by these programs included infrastructure
cold-chain and conditioning improvements, phytosanitary
improvement programs (especially the control of fruit fly
infestation), certification programs, traceability programs,
training in orchard management practices, and postharvest
handling training programs. Figure 10.3 summarizes the
continuum of projects and initiatives that supported the
development of the mango industry in Mali between 1993
and 2009.
To strengthen human and physical capacity and to
improve the competitiveness of mango exports, a pack
house and logistics facility known as the PLAZA (Périmètre
Logistique Aménagé en Zone Aéroportuaire), with capacity
to handle 2,000 tons of fresh produce a year, was built in
2007 (box 10.3). The PLAZA has been integral in helping
mango exporters (mainly sea-freight exporters) prepare
their products for export with precooling and storage
rooms, meet international standards for quality and safety,
and train staff. The PLAZA is the only modern pack house

Figure 10.3 The Stages of Development of the Mango Export Industry in Mali

Mali introduces multi- Mango exports to EU increase

modal transport by Partnership b/w APROFA 600% from 1996–2006
shipping mangoes by sea & Ivorian firm to test the
model API-Mali replaces CNPI New PLAZA begins
APROFA established operations
New investment code adopted
10
8
6
Tons

4
exports in MT
Malian mango

2
0
93

96

98

00

01

03

05

06

07

09
19

19

19

20

20

20

20

20

20

20
PAVCOPA PCDA

World Bank World Bank

DHV Trade MALI
USAID USAID
CAE
IICEM
USAID
USAID

Source: J. E. Austin Associates.

Note: DHV, API and CNPI are all project acronyms.

Box 10.3 The PLAZA

The World Bank project, PCDA, with funding from the
Dutch Embassy, built a modern pack house in Bamako,
near the airport, to help exporters improve their capaci-
ties in handling and shipping mangoes. The PLAZA has
proven to be an effective pack house despite its location,
Source: Authors.
away from the growing region and the border with Côte
d’Ivoire. However, the PLAZA currently operates only
during the mango season. PCDA is looking at other per-
ishable export products, such as papaya and other fresh
produce, that could make use of the pack house.

174 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
for horticulture in Mali. A precooling and shipping facility
in Sikasso is also being considered.
With assistance from the donor community, the govern-
ment has provided technical assistance and training in phy-
tosanitary issues. Continuous tightening of the EU food
safety regulations, coupled with stricter demands in terms
of traceability, are making it necessary to devise and deliver
substantial interventions in agricultural export countries.
Figure 10.4 illustrates the general stages of horticulture
export quality requirements.
Phytosanitary requirements are still a challenge for
Malian produce and several other West African countries.
In the past, many shipments of mangoes were rejected
once they reached Europe because of fruit fly infestations
or other phytosanitary concerns. These problems arise
partly from a lack of proper harvest and postharvest
handling techniques. Since the fruit fly problem is a
regional one, several regional initiatives have been estab-
lished to eliminate the flies. Over the past three years, the
West African Fruit Fly Initiative (WAFFI), jointly financed
by the World Bank, the European Union, and the World
Trade Organization (WTO), has been piloting fruit fly
surveillance and mitigation protocols in West African
orchards, including those in Mali. The knowledge and
experience gained thus far has led to the design of a West
African Regional Action Plan to control fruit flies. Once
the substantial funding required to implement this
program (€25 million) is gathered, the large-scale inter-
ventions planned by the program should translate
into much-reduced prevalence of fruit flies in Malian
orchards, thus improving the quality of the marketed
fruit.

Figure 10.4 Continuum of Commercial and Regulatory Horticulture Export Requirements

+ specifications
for more
+ specifications advanced/quite
for more specific process
advanced and standards, yet
+ specifications implemented in
often quite
for selected, the context of
specific process
basic standards, highly
standards with
basic GAP, good integrated
associated
+ internal hygiene and supply chains
greater detail in
Upgrading steps

quality approaches to and where the

record-keeping
characteristics safe pesticicde supplier has a
of products use/storage and relatively
associated sophisticated
+ quality + basic record-keeping management
grades and requirements systems structure for
varietals on pesticide use
quality control
preferences and risk
+ visual
+ consistent management
characteristics
quality and
quantities 1st and 2nd + 2nd and 3rd
party + 2nd and 3rd party conformity + 2nd and 3rd
inspections/ party conformity assessment party conformity
+ visual + visual assessment
inspections inspection testing assessment

Level 1 Level 2 Level 3 Level 4 Level 5 Level 6

Stringency of official and buyer’s requirements

Level of sophistication of the conformity assessment systems

Informal sector - Domestic high end horticulture Horticulture exports to high-end Int’l outlets
domestic - Ethnic produce exports Low risk products High risk
- Exports to wholesale markets products/sophisticate
(fresh and dried products) postharvest operations

Source: Jaffee 2009.

CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 175
 Phytosanitary issues are closely linked to the certification
process. Recognizing that many of its mangoes are sold in
supermarket chains in Europe, Mali has pursued several cer-
tifications, including GLOBAL G.A.P. certifications. Stream-
lining the export process at the airport has contributed to
the success of Mali’s certification programs.
Improving orchard management and
postharvest handling interventions
Orchard management and postharvest handling interven-
tions have been very important in Mali as well. To have
high-quality fruit to export, quality improvement must
begin at the orchard level. Training in such improvements
has been provided by several institutions in recent years,
including the government, international partners, and
even exporters themselves. Training has been comprehen-
sive, with topics ranging from harvesting best practices,
transport handling aimed at reducing damage to the
product, and grafting techniques to improve the varieties
and productivity of trees. Over time, training has
increased the capacity of mango producers, allowing for
continued success in the subsector.
Supply chain financing
Through the PCDA project, local banks and other financial
institutions have regained trust and interest in Malian hor-
ticulture. In 2008 CFAF 150 million were lent to operators
in the subsector, a record level compared with the past 10
years. This loan was disbursed in the form of credit for the
import of shipping material (CFAF 56 million) and seasonal
credit (CFAF 94 million) for three exporters. Between 2007
and 2008, PCDA provided a guarantee to a commercial
bank to fund the import of transport boxes from Côte
d’Ivoire; the bank held them in bond through a third
party, releasing them on credit terms when the exporters
needed to prepare an export shipment. In 2009 the same
banks agreed to pre-finance the import of boxes directly to
the exporters. PCDA has established its credibility in this
area and continues to play an important facilitative role
between farmers, exporters, and professional operators on
the one hand, and the banks and financial institutions on
the other hand.
Business training
Capacity building has been critical for the transition
from subsistence mango farming to commercial farming.
Producers received training in crop husbandry and best
176 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
harvest practices, developing tools and linking value
chain actors to financing. International partners have
also provided assistance in small business development
techniques.
As part of PCDA, for example, the Dutch firm Bakker
Barendrecht teamed with five exporters operating at the
PLAZA to teach the exporters how to reach European mar-
kets with their products. Bakker invests in knowledge in
mango production and pays a minimum on the fruit if the
exporters abide by a code of practice detailed in the contract.
Varieties, ripeness, size, and other details are agreed upon
with exporters for the entire season, with the possibility of
payment of premiums if the mangoes sell well.
RESULTS AND IMPACTS
Mali’s development agenda places a high priority on agri-
cultural growth and diversification. The achievements
realized through the PAVCOPA and PCDA projects have
made a direct and tangible contribution to this agenda. A
whole range of stakeholders participating in the value
chain—small farmers, traders, agroprocessors, exporters,
service providers (technicians, financiers, and accounting
specialists), and input and equipment providers—have
been involved since project launch and are benefiting
from the expansion and improvements brought about by
these projects.
Quantitative results
Results achieved after 15 years of reform in Mali’s mango
subsector are summarized in table 10.3. In many cases, the
improvements are significant. The volume of mangoes
exported, for example, reached 11,995 tons in 2008, an
all-time record with a growth rate of 24 percent a year.
(figure 10.5). Mango exports generated revenue of CFAF
9.7 billion ($25 million) in the same year, a significant
proportion of the earnings generated by Mali’s traditional
exports, such as cotton lint. A range of stakeholders—
farmers, harvesters, processors, and exporters—are
benefiting from expansion in mango trade and improve-
ments along the value chain. This means increased market
share, value creation, and improved prices at all stages of
mango production. The price producers receive for their
mangoes at the farm gate increased by 150 percent
between 1993 and 2008. That price increase has allowed
for additional capacity for other on-farm activities as
farmers became aware of the margins to be gained from
horticulture production.
 Table 10.3 Key Quantitative Improvements in Mali’s Mango Subsector, 1993–2008
1993 2008 Impact
1,050 metric tons exported 11,995 tons exported 1,042%
Marginal sea-freighted mango exports; exports not recorded as Sea-freighted exports total 4,600 tons 460%
originating in Mali
$460,000 of revenue generated by mango exports Value of mango exports reached $3.4 million in 2007 639%
25-day transit time from Sikasso to northern Europe 12-day transit time from Sikasso to northern Europe –13 days
Farm gate price of CFAF 50 Farm gate price of CFAF 125 150%
237 tons of European imports European imports increased to 4,560 tons 1,824%

Source: Author research; revenue contribution figures from FAOSTAT.

Figure 10.5 Mali Mango Exports, 1993–2008

9
8
7
6
5
4
3
2
1
0
93

94

95

96

97

98

99

00

01

02

03

04

05

06

07

08
19

19

19

19

19

19

19

20

20

20

20

20

20

20

20

20
Mango export quantity (million metric tons) Mango export value (million US$)

Source: FAOSTAT and J. E. Austin.

The average quality of fruit exported from Mali has

improved. The number of sea container rejections due to
fruit flies, for example, dropped from 14 containers in
2007 to 5 containers in 2008. Backward linkages at the
production level also have improved tremendously, as
have relationships between exporters and farmers. Mali
has also been able to set a foothold on the Fair Trade
niche market (box 10.4). Exporters provide support ser-
vices to farmers such as helping to manage their planta-
tions, working to reduce fruit flies, and implementing
certification or traceability programs on the plantation.
In return, exporters purchase farmers’ final product—
often without a contract. The trust established by these
interactions allows exporters to obtain a higher-quality
product, because farmers are more willing to respect phy-
tosanitary controls when provided assistance. In fact,
trust among all actors in the mango value chain has
increased over time.
Box 10.4 Fair Trade Market Opportunities
Recent data from the Fairtrade Labelling Organiza-
tions show that about 8,000 tons of fair-trade-certified
fruit were sold in Europe in 2005, including pineap-
ples, mangoes, avocados, and citrus and deciduous
fruits (excluding bananas). By volume, the United
Kingdom accounted for more sales than any other
country, with 4,700 tons of fruit. On a per capita basis,
however, spending was highest in Switzerland. Sales of
fair-trade products have grown strongly since 2005.
As a brand, fair trade is making its way into the
mainstream and is being taken up by supermarkets.
The sector already has the same performance require-
ments as conventional products: year-round supply
and quality and price guarantees are to be expected.
Source: Authors.

CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 177
 In addition to improvements within Mali, regional
cooperation also has improved. For example, mango
exporters in Mali have organized to obtain multicountry
support from donors to fight against fruit flies, and the
multimodal transport initiative involved coordination
among several stakeholders across borders. These efforts
have led to an improved value chain not only in Mali but in
Côte d’Ivoire, Senegal, and other countries.
Qualitative changes
The initial, and now sustained, take-off in the growth of
mango exports has led to a complete transformation of the
subsector, not only in quantitative terms but also qualita-
tively. The expansion of the subsector has brought about a
progressive and likely irreversible change in business prac-
tices in the sense of increased professionalism and attention
to product quality, better compliance with trade standards,
and increasing interest in private investment.
For one, major players in the mango business are now
present in Mali and have an interest in expanding their
operations and making long-term investments in the coun-
try. AHOLD, a major Dutch supermarket retail chain, has
operated in Mali since the PLAZA and has been providing
technical assistance to Malian exporters and PCDA. In
addition, substantial work has been undertaken on the
upstream/production level of the production chain. Mango
production is now considered a legitimate agricultural
activity, not just fruit collection. (PCDA, for example, has
invested in a study that aims to map tree crop plantations
using satellite imaging.) Financial institutions have also
shown renewed interest as demonstrated by the increasing
volume of credit to the subsector, the low default rates, and
the emergence of innovative financing instruments, such as
input prefinancing. Finally, Mali’s Mango Task Force has
been active for some years to improve coordination in the
subsector, develop an agenda for collective action between
professionals, and provide a platform to address issues of
common interest with the public sector.
THE WAY FORWARD
While the mango industry in Mali has grown in recent
years, it will continue to develop in the years ahead. Mali has
given priority to a number of areas, including market posi-
tioning, improving quality and quantity, attracting capital,
adding value to mango exports, improving transportation
infrastructure, and potentially establishing more links at the
regional level.
178 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
Improving market positioning and
diversifying market outlets
While Mali has succeeded in increasing sea-freighted exports
since the 1990s, there is still room for better market posi-
tioning and diversification of market outlets. First, Mali,
whose mangoes represent only 2 percent of the total mango
imports to the European Union, is still a marginal player in
the global market for mangoes (figure 10.6). Second, the
European Union and other importing markets are dynamic,
meaning that there is constant need for adjustment and
improvement to meet market demand. Mali is currently
supplying the EU market based on the seasonality of its pro-
duction, which is concentrated between mid-May and early
July, after Burkina Faso but before Senegal (figure 10.7).
Going forward, it will be important for Mali to extend
its harvest season and develop markets other than Europe.
To address this, PCDA is testing the irrigation of mango
orchards, which may offer opportunities to induce
flowering and fructification at a different period of the
year. Local exporters also have begun seeking market
opportunities in the Middle East and North Africa,
among other places. Mali has begun to export mangoes to
Morocco. There are also substantial opportunities in
other Sub-Saharan country markets.
Continuing to invest in quality and
product differentiation
In recent years Mali has focused on obtaining certifications
that will allow mangoes to enter the European market, such
Figure 10.6 Volume of Mango Imports to the European
Union, by Country of Origin, 2008
Other
Costa Rica Mali
2%
10%
2%
Senegal
3%
USA
3%
Cote Brazil
42%
d'Ivoire
5%
Israel
5% Peru
22%
Pakistan
6%
Source: Eurostat.
Figure 10.7 Seasonality of Mango Imports to the European Union, by Country of Origin, 2008
(metric tons)
25,000
20,000
15,000
10,000
5,000
0
1 2 3 4
Others Mali
Israel Pakistan
Source: Eurostat.
as GLOBAL G.A.P. the certification required by supermar-
kets in the European Union. Collaboration along the value
chain led exporters to assist producers to achieve certifica-
tion through training, with the support of PCDA. Some
producers have even obtained certification of their man-
goes as organic. The market for organic fruit and vegeta-
bles is a growing one, particularly in countries of northern
Europe and Switzerland.
Ensure compliance with standards and
phytosanitary requirements
The constant evolution of EU food safety regulations toward
more stringent phytosanitary requirements, coupled with
demands for traceability, make it necessary to devise and
deliver specific interventions in the produce-exporting coun-
tries to meet market requirements. Recently PCDA financed
an intervention by the Comité de liaison EU-ACP pour les
fruits et légumes (COLEACP) specific to the mango value
chain aimed at building capacities in health and phytosani-
tary risks control. A series of workshops, which brought
together mango stakeholders of Mali and Burkina Faso, were
held to train participants on risk assessment and mitigation.
This training led to the creation of a guidebook for the mango
value chain that is now followed in both countries.
Continuing to focus on improvements
in quantity and quality
Although estimates vary greatly, some experts estimate that
Mali could export between 20,000 and 50,000 metric tons
CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
5 6 7 8 9 10 11 12
Months
Senegal Costa Rica USA
Cote d'lvoire Peru Brazil
of mangoes a year with improved management of existing
mango plantations. Effective management includes intro-
duction of new varieties, improved grafting and replanting
of new trees for increased yields and quality, additional
investments in pack houses and cold storage, renewed
efforts in the control of fruit flies, and procurement of tech-
nical advisory services. However, in 2009, against an initial
export program of 104 containers, the PLAZA could finally
ship only 42 containers because of the lack of mangoes of
exportable quality.
Attracting capital
Developing commercial value chains is a private sector busi-
ness. It requires capital and know-how. If an economic sec-
tor does not manage to attract investors and capital, both
national and foreign, sustaining growth will be difficult. In
general, governments should encourage commercial farm-
ing by facilitating access to land concessions. In Mali’s
mango subsector, the government should make financing
more readily accessible and provide incentives to banks to
lend to stakeholders. These actions would stimulate private
sector investment into value-addition activities such as dry-
ing, canning, and juice production. Burkina Faso has had
some success in producing dried mangoes, and Mali has the
potential to achieve similar success.
Diversify to better utilize the PLAZA terminal
To maximize the use of the PLAZA and ensure sustainabil-
ity, commercial operations must be able to cover PLAZA
179
 operating costs by the end of 2012, when the public funding
stops. Several exporters are looking at opportunities to
export green beans and other vegetables that could be
packed at the terminal. To ensure profitability and com-
petitive pricing for its members, there are plans for a
strong professional organization to eventually take over
management of the PLAZA.
Further invest in marketing infrastructure
A mechanical grading line is being installed to improve
turnover at the PLAZA, enhancing the ability to quickly
meet demand quantity and quality in the European
market. The equipment was ordered by PCDA, and the
calibrating machine should be installed for the 2010 sea-
son. Assuming an adequate fruit harvest, the PLAZA
should now be able to handle over 100 containers in the
12- to 14-week mango season.
Adding value to consolidate the achievements
Strengthening and diversifying the value chain, in addition
to valorizing the large mango surplus that is not exportable,
is critical for the future growth of the mango subsector.
While fresh mango exports were an entry point for Mali to
reach the EU market (and indeed, demand for fresh man-
goes is expected to continue in the foreseeable future), fresh
mangoes may not be extremely profitable in the long run,
and Malian professionals need to seek different ways to add
value in the mango subsector. PCDA has been working on
several clear opportunities: processing, increasing the range
of products offered, expanding other value chains (papaya),
and establishing new value chains.
A recent study (Royal Tropical Institute 2010) shows the
potential for exports of dried mango from Mali to European
markets, provided the product meets the expected quality
standards. For dried mangoes, Mali has an advantage over
Burkina Faso in the sense that it is relatively unencumbered
by an obsolete technology and enjoys a large surplus of
mango varieties that are not exportable as fresh. Recently,
PCDA agreed to take up one of the study’s recommenda-
tions and finance a pilot project to produce dried mangoes
based on technology established in South Africa, which cur-
rently meets the quality standards in EU markets.
Further improving transport and
fostering cluster development
The inefficiencies the Bamako-Dakar railway and at the
Dakar port are resulting in higher costs, longer wait times,
180 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
and higher risks as Malian exporters operate through Côte
d’Ivoire. Improvements in alternative transport routes
would certainly benefit the mango subsector. Mali could
also improve the road-to-rail corridor in order to reduce
transport times and increase quality.
To develop a vibrant and competitive export horticulture
sector, Mali also needs to encourage the development of
related industries and services, such as certification, packag-
ing material, and inputs. These related industries are key to
improving the competitiveness of the value chain(s) by
reducing the cost of inputs, technology, and services, and
also by providing an important source of revenue and
employment in Mali.
Joining forces at the regional level
Finally, when examining trade figures, it is evident that the
competition for West African mango sectors does not come
from neighboring countries, but from Central and South
America. The concentration of mango exporters within the
West African region, however, suggests advantages from
coordinating the production and export of mangoes to
Europe at a subregional level. Mali could reap significant
advantages from coordinating mango production and
export with Côte d’Ivoire, Burkina Faso, and Senegal. Such
coordination could result in an increase in the scale of
export and greater efficiencies in shipping and standardized
quality control, better handling of traceability, and brand-
ing. All these factors would contribute to increased volumes
of exported mangoes and, ultimately, to increased revenues
for producers. One possibility would be to create a West
Africa brand for mangoes.
LESSONS AND REPLICABILITY
The driving forces behind Mali’s experience in developing
its mango sector, as well as the lessons learned during the
development process, are useful to point out for other coun-
tries in the course of such a transition.
Drivers of success
Innovation Is Critical in Triggering and Driving
Change. In the case of mango exports from Mali, the key
initial innovation was in transport and logistics. Together,
these changes created greater opportunity to access the large
EU market. The innovations also created the dynamics of
change and initiated a learning process. There is no univer-
sal formula for achieving innovation, however; innovation
derives from a combination of factors that depend on the
specific country and sector context. In the case of the mango
subsector of Mali, two key resources, expertise and entre-
preneurship, were brought together to improve the export
capacity of the industry. Additionally, creativity in funding
mechanisms is another important lesson learned in the case
of Mali.
Time Is of the Essence. Building capacity—human, phys-
ical, or otherwise—in a new or small industry requires sus-
tained efforts over time, especially when starting from
scratch or from a very low base, as was the case in Mali.
High-Quality Technical Work Is Key. The importance
of market research, value chain cost analysis, benchmarking,
and assessment of industry constraints cannot be over-
stated. This work is critical in identifying and designing
action plans, programs, and business solutions. In the case
of Mali, it was perhaps the most critical factor contributing
to launching the reforms.
Private Sector Leadership Is Also Important. Even
though private sector involvement in the mango sector in
Mali was weak at the start of the transition, it was neces-
sary to work with the existing private operators and even-
tually bring in new ones, such as the Ivoirian company
that conducted the pilot export test. In the case of Mali,
the partnership with SN Tropical Expressions was unique
and provided a good model of what public-private part-
nerships can achieve.
Knowledge and Funding Must Be Packaged and
Delivered Properly. Building capacities in emerging
subsectors and industries takes not only time and perse-
verance but also the ability to deliver investments in
“hardware” (infrastructure and other means of produc-
tion) and “software” (training and knowledge transfer) in
a flexible manner. In the case of Mali, capital investments
such as PLAZA were imperative to the success of the
mango subsector. Training alone could not bring about
such a transition. Improvements to cold chain, trans-
portation, and conditioning facilities within the horticul-
ture sector significantly improved Mali’s capacity to
export quality mangoes.
Lessons learned
A Targeted, Organized Approach Is Necessary to
Sector Reform. Sector and value-chain analysis are essential
CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
tools in identifying opportunities and articulating opera-
tional strategies to create greater value in agriculture and
agribusiness. It is also essential to find a key entry point that
responds to market demand and helps increase scale. For
Malian mangoes, this meant finding a more effective way to
get the product to the market, while innovations in trans-
port and logistics system allowed farmers and exporters to
achieve the economies of scale they needed. A structured,
holistic approach is required.
Sustaining Development Efforts over Time Is Neces-
sary. Mali could not have successfully transformed its
mango sector if activities had not been pursued year after
year. Building an industry takes years, not months. The les-
son here for governments and development partners is that
if they not prepared to dedicate time and resources to a
reform over a long period of time, it is probably better not
to start at all. That said, early success in the reform process,
such as what Mali achieved through the positive outcome of
the multimodal export pilot launched by APROFA in 2000,
can serve as a foundation on which to build momentum for
further reforms.
Aid Funds Can Play a Catalytic Role in Change
Processes. Project aid has the capacity to provide both
financial resources and know-how to share risks in order to
facilitate innovation. In Mali, technical assistance programs
targeting postharvest handling, supply chain finance,
export, and other areas have helped improve the perfor-
mance of the mango value chain and are now doing the
same for other agricultural and nonagricultural sectors.
Technical assistance by aid organizations can also help
address market failures and can be combined in a proactive
fashion with efforts and interventions by the private sector.
Collaboration and Partnerships Are Part of the
Solution. Public-private partnerships (between donor
agencies and government, or between the private sector and
government), such as the arrangement to build and manage
the PLAZA, are key in developing infrastructure that will, in
turn, help a country expand a specific sector. In Mali the
emergence of the Mango Task Force helped provide a space
for public-private cooperation, as well as an opportunity for
exporters to coordinate their respective shipments through
a common buyer. In addition, the use of “Mali mango” logo
on boxes created closer cooperation among shippers.
A Favorable Investment Climate and Investment
Policies Remain Important. Thus far the government of
181
 Mali’s commitment to reform in its mango sector has been
carried out through diversification, donor financing, and
improved technical expertise. The industry, however, is now
experiencing a new generation of issues that will require sig-
nificant reforms in the general business environment by
promoting private sector investment, and improving access
to land. Recent difficulties in implementing a Foreign
Investment Advisory Service (FIAS) technical assistance
program on agribusiness do not bode well in terms of the
determination of the government to move to the next stage
of development of agribusiness value chains.
Scalability and transferability
Two general lessons from Mali’s experience with mangoes
can be applied to other countries. First, designing market-
led strategies and investing in applied value-chain analysis is
necessary if the objectives of creating higher value in agri-
culture, raising incomes, and lifting the rural population out
of poverty are to be achieved. While general, cross-cutting
agricultural support programs are necessary, they are in
themselves not sufficient to make a difference. Specific
information and analysis must be generated on the issues to
be addressed throughout the value chain(s).
Second, it should be kept in mind that sustainable mar-
ket inclusion requires multiple interventions. As demon-
strated by a recently completed research program funded by
the U.K. Department for International Development called
“Re-governing Markets,” inclusion of small farmers in mod-
ern agricultural value chains does not happen by itself. In
fact, inclusion requires interventions by multiple actors, as
summarized in figure 10.8. The study indicates that three
major categories of actors need to be involved: farmers and
Figure 10.8 Multiple Actors Necessary for Sustainable
Markets
Multiple interventions
Receptive business Trained/
Partnership
sector
facilitation
empowered
farmers
Facilitating public
sector
Source: UK Department for International Development.
182 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
farmer groups, receptive businesses, and a facilitating public
sector. Actors in all categories must be willing and able to
play their role, and even then establishing sustainable mar-
ket inclusion needs a public sector body to facilitate the
process. This is the role that APROFA and its successive
projects have played for the mango subsector in Mali and
the role that such projects are still playing for a number of
other emerging value chains in the country (papaya, cashew,
sesame, potato, and onion).
CONCLUSION
Since the early 1990s Mali has achieved a spectacular
increase in its exports of fresh mangoes, seizing opportuni-
ties offered by an increasing market demand in Europe and
finding solutions to overcome significant physical hurdles.
The initiatives taken in the mango sector in Mali clearly
demonstrate that even a physical constraint as immutable as
being landlocked can be overcome to some extent through
innovative solutions. In Mali the entry point was the identi-
fication and economic validation of a new transport and
logistics arrangement that allowed mangoes to be exported
in large volumes and in good condition.
Mali currently is focusing on consolidating and expand-
ing its initial success by strengthening stakeholder involve-
ment and private sector partnerships to ensure growth and
sustainability. The Mango Task Force has emerged as the
sector’s professional organization, with potential to become
a permanent trade/business association capable of handling
the new issues facing the horticulture industry in Mali.
BIBLIOGRAPHY
Antoine, M. 2005. “Growing Malian Mango Exports: How
New Business Models Translate to Increased Opportuni-
ties for West African Producers.” World Bank.
APROFA (Agricultural Value Chain Promotion Agency).
2002a. “Mango Exports to Europe out of Sikasso-Mali:
Multimodal Shipping and Cross-border Partnerships.”
Bamako.
———. 2002b. “Note sur l’expérience de Sikasso concer-
nant l’amélioration des revenus des petits planteurs de
mangues. ” Bamako.
———. 2002c. “Projet de diversification des filières agri-
coles et de l’horticulture : Livrable 1.1: Diagnostic des
principales filières agricoles (mangue, haricot vert,
pomme de terre et oignon).” Bamako.
Birthal, P., A. Jha, and H. Singh. 2007. “Linking Farmers to
Markets for High-Value Agricultural Commodities.”
Agriculture Economics Research Review 20: 425–39.
CARANA Corporation. 2004. “Impact of Transport and Logis-
tics on Mali’s Trade Competitiveness.” Washington, DC.
Chemonics International. 2003. “Mali SEG Projet Centre
Agro-Entreprise: Final Report Deliverable 4.” Washington,
DC.
Club du Sahel/OECD (Organisation for Economic Co-
operation and Development). 1998. “L’économie locale
de Sikasso.” Paris.
Danielou, M., P. Labaste, and J.-M. Voisard. 2003. “Good
Practice Note: Linking Farmers to Markets: Exporting
Malian Mangoes to Europe.” Africa Region Working
Paper 60, World Bank, Washington, DC.
IF (Integrated Framework for Trade-Related Technical
Assistance to Least-Developed Countries). 2004.
“Expanding and Diversifying Trade for Growth and
Poverty Reduction: A Diagnostic Trade Integration
Study.” http://www.integratedframework.org/files/mali
_dtis_17nov04.pdf.
Ezzine, A. 2009. “Etude des corridors pour l’export des
mangues du Mali et du Burkina Faso vers l’Union
Européenne.” World Bank, Washington, DC.
Fairbanks, M., and S. Linsay. 1997. “Plowing the Sea: Nur-
turing the Hidden Sources of Growth in the Developing
World.” Harvard Business School Press, Cambridge, MA.
Geomar International and PCDA. 2006. “Strategic Profile of
Mali’s Agricultural Exports.” Geomar International,
Montreal.
Gulati, A., N. Minot, C. Delgado, and S. Bora. 2007. “Growth
in High-Value Agriculture in Asia and the Emergence of
Vertical Links with Farmers.” In Global Supply Chains,
Standards and the Poor, ed. J. Swinnen, pp 91–108.
Wallingford: CABI Publishing.
Hallam, D., et al. 2004. “The Market for Non-Traditional
Agricultural Exports.” Food and Agriculture Organiza-
tion, Rome.
Jaffee, S. 2009. “African Smallholders, Higher Value Agri-
Food Markets and Measures to Achieve Compliance with
Emerging Standards.” World Bank, Washington, DC.
Jaffee, S., and S. Henson. 2004. “Standards and Agro-Food
Exports from Developing Countries: Rebalancing the
CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN
Debate.” Policy Research Working Paper 3348, World
Bank, Washington, DC.
Jaffee, S., P. Labaste, R. Kopicki, and I. Christie. 2003. “Mod-
ernizing Africa’s Agro-Food Systems: Analytical Frame-
work and Implications for Operations.” Africa Region
Working Paper 44, World Bank, Washington, DC.
Jones, C., and G. Livingston. 2009. “Supporting Agribusi-
ness Investment in Mali.” J. E. Austin Associates,
Arlington, VA.
PCDA (Programme Compétitivité Diversification Agri-
coles). 2007. “Mango Value Chain Competitiveness
Plan.” Bamako.
Rey, J.-Y., et al. 2004. “La mangue en Afrique de L’Ouest
Francophone: variétés et composition variétale des verg-
ers.” Fruits 59 (3): 191–208.
Royal Tropical Institute. 2010. “L’amélioration des perform-
ances de la filière des produits séchés de la mangue au
Burkina Faso et au Mali.” Amsterdam.
USAID (U.S. Agency for International Development). 2002.
“Mali Trade Capacity Needs Assessment—Sub Sector
Analysis.” Technical Report, Washington, DC.
World Bank. 1995. “Staff Appraisal Report, Agricultural
Trading and Processing Promotion Pilot Project.”
Washington, DC.
———. 1996. “Mali Agricultural Trading and Processing
Promotion Pilot Project.” Project Appraisal Document
25884, World Bank, Washington, DC.
———. 2003a. “Implementation Completion Report
Republic of Mali Agricultural Trading and Processing
Promotion Pilot Project.” Washington, DC.
———. 2003b. “Mali: Exporting Mangoes to Europe.”
Washington, DC.
———. 2004. “Project Appraisal Document on a Credit and
Proposed Grant to the Government of Mali for the
Transport Corridors Improvement Project.” Washington,
DC.
———. 2005. “Project Appraisal Document on a Proposed
Credit to the Republic of Mali for an Agricultural Com-
petitiveness and Diversification Project.” Washington,
DC.
183
 CHAPTER 11
Economic Liberalization in Rwanda’s Coffee
Sector: A Better Brew for Success
Karol C. Boudreaux
n 2008 Solberg and Hansen, a Norwegian importer of
I
high-quality specialty coffee, bid just under $40 a kilo-
gram (more than 21,000 Rwandan francs) for a load of
coffee from Rwanda.1 While this bid was exceptional—
specialty coffee normally sells for between $3 and $4 a
kilogram—it also represents a real and positive transfor-
mation within Rwanda’s coffee sector. In the not-too-
distant past, the country was known as a producer of
mediocre coffee that attracted little attention from dis-
criminating importers or consumers.2 Today, Rwandan
coffee is increasingly recognized as a high-quality prod-
uct, one for which importers such as Solberg and Hansen
and, in turn, consumers are willing to pay a premium.
Although the Rwandan economy is diversifying, agricul-
ture continues to be the primary source of livelihood for
the vast majority of the population. The overwhelming
majority of agricultural workers are subsistence farmers.
Just over 10 percent of these farmers planted coffee in
2008, and the crop remains a major source of export rev-
enue for Rwanda, generating more than 36 percent of
total export revenue in 2009.
The evolution of coffee production in Rwanda
The transformation of Rwanda’s coffee sector happened rel-
atively quickly. In 2000 Rwandan farmers were producing
semi-processed coffee for sale on world markets. Farm gate
prices paid to farmers were low (60 Rwandan francs a kilo-
gram), and prospects for farmers and exporters to increase
income or profits were limited. Starting in the late 1990s, the
government liberalized the sector, removing a variety of
barriers to trade, creating new incentives for groups and
individuals to invest in coffee production, and facilitating
entrepreneurship in the coffee industry.
Working with the private sector and international
donors, the government of Rwanda has reshaped the coffee
industry: the regulatory framework for production has been
modified; more than 100 coffee-washing stations have been
built; donors have supported the development of market
links between producers and foreign buyers; cooperatives
have formed; and smallholder farmers are working together
in an effort to improve the quality, marketing, and branding
of their coffee.
These changes have had important effects on the ground
in Rwanda. Coffee continues to generate important export
revenue for the country: just over $47 million in 2008, com-
pared to $35 million in 2007. Higher incomes benefit farm-
ers, their families, and their communities in a variety of
ways: farmers can improve a home, pay medical expenses
or school fees, or better ensure food security. And when
cooperatives earn a profit they are able to hire workers,
purchase capital, and support community projects such as
improved schools.
The transformation of the Rwandan coffee sector may be
creating social benefits in addition to economic benefits.
Working together at coffee-washing stations—typically
located in the rural, hilly, and relatively inaccessible areas
where coffee grows and where little other commercial
185
 infrastructure exists—farmers have new opportunities to
interact with other Rwandans. These repeated interactions
may reduce the sense of ethnic distance among members
of Rwandan society. As farmers and other workers at
coffee-washing stations experience the increased eco-
nomic satisfaction that comes with higher income earned
from coffee, they may also feel greater levels of trust
toward people with whom they interact and thus develop
more positive attitudes toward reconciliation. And because
coffee in Rwanda is grown by poor, smallholder farmers
who make up 90 percent of the population, these positive
changes have the potential to benefit a broad swath of
Rwandan society.
Challenges and concerns
Despite this good progress, the coffee sector faces a number
of serious challenges. To start, Rwanda’s landlocked geo-
graphical status and poor infrastructure mean that coffee
producers face high transport costs. Moving coffee cherries
quickly over Rwandan roads is difficult, as is moving
processed beans out of the country in a timely and cost-
effective manner. Other concerns are related to the costs in
the industry. Production costs remain high. Many of the
newly built coffee-washing stations are operating at much
less than full capacity, and labor costs are higher than in
neighboring countries. Rwandan farmers are also less pro-
ductive than farmers in neighboring countries. Although
some have received support and training from nongovern-
mental organizations (NGOs), regular visits from extension
agents are limited.
In addition, a variety of management concerns have
plagued the cooperatives that many coffee farmers join.
One specialist of Rwanda’s cooperatives observed, “After
five years of extensive cooperative capacity building,
Rwanda’s coffee cooperatives remain surprisingly fragile,
unorganized, and dysfunctional” (SPREAD 2007). Some
cooperatives have mishandled loans, while others have failed
to fulfill contracts in a timely manner or have encountered
trouble marketing their products. Some of these problems
are the result of a lack of training or financial management
skills.
Other challenges involve the broader institutional envi-
ronment. As Rwanda implements a new land law, some
smallholder farmers may face uncertainty about the rights
to the land that they work. Women, in particular, may be
especially vulnerable. The government, NGOs, and other
stakeholders are taking measures to deal with these various
challenges, however. If capacity issues can be addressed,
186 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
marketing and sales problems resolved, incentives
strengthened to produce higher-quality beans, and harm-
ful government interference avoided, then the positive
gains of the past several years should continue, and
Rwanda’s smallholder farmers can look forward to earning
more income from coffee production. This income should,
in turn, filter through local economies to spread benefits
to other Rwandans.
REFORMING A VITAL SECTOR
Government control and limited markets
In many developing countries, governments are heavily
involved in the agriculture sector, and that certainly holds
true in Rwanda, where coffee has been a major export for
decades. The Belgian government as well as the two inde-
pendent, pre-genocide governments controlled important
aspects of the coffee trade for their political and financial
gain. Through compulsory production, export taxes, and a
monopsony export control agency, these regimes captured
the profits of mostly poor coffee farmers, and used the
funds to help maintain political power (Bates 1981). Pro-
ducers had little incentive to invest in the production of
high-quality coffee, and so for decades Rwandans produced
a small volume of low-quality coffee.
Significant government involvement in Rwanda’s coffee
sector began in the 1930s, when the Belgian colonial gov-
ernment launched a series of “coffee campaigns.” Govern-
ment authorities built nurseries and supplied seeds, but
they also required Rwandan farmers to plant coffee trees
(Dorsey 1983). The government also introduced price
restrictions, imposed mandatory quality guidelines, and
issued special licenses that allowed only some firms to pur-
chase coffee. Export taxes were imposed on coffee sales, and
individual income taxes were imposed on producers, who
at the time were mostly Hutu farmers. Tutsi chiefs collected
these taxes, which helped support them and the colonial
government.
Following Rwanda’s independence, the Kayibanda gov-
ernment (1962–73) retained most of these policies because
it had limited alternatives for raising revenue. A government
marketing board (OCIR, subsequently OCIR-Café),
together with a monopsony export company, Rwandex,
purchased, and then sold on world markets, the vast majority
of coffee grown in Rwanda. The farm gate price was set by
the government.3 Middlemen bought beans from farmers
and sold them to Rwandex, which in turn sold them to
foreign buyers. The locations where smallholders brought
their beans for purchase acted as “the economic arm of the
Gitarama (that is, Kayibanda) regime” (Verwimp 2003).
Heavy government involvement in the coffee sector con-
tinued under the Habyarimana regime (1973–94). During
the 1970s and 1980s, as world coffee prices rose, coffee
exports provided between 60 and 80 percent of Rwanda’s
export revenue (Berlage, Capéau, and Verwimp 2004).
President Habyarimana ensured control of these important
rents by appointing relatives and political supporters to
positions of authority at OCIR-Café.4
Crisis and response
During the early and mid-1980s, rising coffee prices allowed
the government to modestly increase the price it paid to
farmers for their beans. Verwimp (2003) indicates that
high market prices for coffee had another effect—they
“allowed the [Habyarimana] regime’s elite to increase both
its personal consumption and its power over the popula-
tion” (p. 172). The government then used the additional
revenue to buy support in rural areas (through higher farm
gate prices for coffee and subsidized agricultural inputs)
and to spend more on monitoring the population.
Tumbling coffee prices in the late 1980s meant reduced
government revenue. For a few years, the government
attempted to keep payments to farmers stable, but this policy
proved unsustainable, especially after 1990, when the govern-
ment needed resources to fight the invading Rwandan Patri-
otic Front (RPF) forces. By the early 1990s the government
was forced to lower prices paid for coffee cherries to small-
holder farmers. Price supports ended completely in 1992.
The fall in coffee prices, coupled with growing military
expenditures, meant that the Habyarimana government
faced severe fiscal constraints. Seeking alternate sources of
revenue, it began confiscating property and raising taxes to
supplement the budget, and there was some amount of
reduction in consumption by elites. Foreign aid also became
an increasingly important part of Rwanda’s budget. How-
ever, the reliance upon, and competition for, foreign aid cre-
ated serious problems among governing elites. As one
observer puts it, “the various gentlemen’s agreements which
had existed between the competing political clans since the
end of the Kayibanda regime started to melt down as the
resources shrank and internal power struggles intensified”
(Prunier 1995, 47). With its growing dependence on for-
eign aid, and in a bid to remain in power, the government
agreed to an International Monetary Fund (IMF) struc-
tural adjustment program in 1990 that imposed further
hardships on farmers.
CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
To rebuild its popularity, the regime diverted attention
from its own economic policies to the Tutsi/RPF threat
and increased levels of repression within Rwandan soci-
ety. The government demonized the invaders, arguing
that allowing Tutsis into the country would lead to Hutus
having less access to already scarce land, and used the
media to foment ethnic hatred. Most repression was
directed at the Tutsi minority, although some spilled over
to Hutus. The ultimate results were, of course, disastrous. In
the three months between April and June of 1994, approxi-
mately 800,000 people were murdered in Rwanda, mostly
Tutsis but also moderate Hutus and Hutu opponents of the
Habyarimana government.
Liberalization to open markets and
increase opportunity
Today, the Rwandan government is less directly involved in
the coffee sector. Farmers have more choice about what to
grow, to whom to sell their beans, and how to market their
product. Private sector investment in the sector is rising. This
increased openness is part of a larger government effort to
improve economic growth in the country. Rwanda’s Vision
2020 is a strategic plan for economic change. This plan has,
since 2003, provided a guideline for sectoral policy setting
with Rwanda’s ministries (Rwanda Ministry of Finance and
Economic Planning 2000; IFAD 2006).
The goals created by Vision 2020, together with Rwanda’s
Poverty Reduction Strategy Paper (PRSP) and the subse-
quent Economic Development and Poverty Reduction
Strategy (EDPRS), include improving the institutional envi-
ronment to allow for increased private sector development
and infrastructure improvements, focusing on good gover-
nance (including democratization, national reconciliation,
political stability, and security), improving agricultural
productivity, improving human capital through invest-
ments in health and education, creating a service-based
economy with a focus on ICT (information and communi-
cation technology), reducing external support, relying more
on exports, and promoting regional integration.
Some progress has been made toward achieving the Vision
2020 goals. Real GDP growth has been strong for more than
a decade: 10.8 percent over 1996–2000 and 6.4 percent over
2000–06 (Rwanda Ministry of Finance and Economic Plan-
ning 2007), reaching a high of 11.2 percent in 2009 (although
it was estimated to fall to under 5 percent for 2010) (Index
Mundi 2010) With an emphasis on private-sector-led
growth and improvements in the environment for doing
business, the economy is diversifying, merchandise trade
187
 levels are rising, and the service sector is expanding. Partic-
ularly for rural Rwandans, reform in the coffee sector is
playing an important part in helping thousands of farmers
increase their incomes, by creating jobs and providing
opportunities for new skills training. The reform measures
are also strengthening human and social capital and, in the
process, may also be generating valuable social benefits.
Most Rwandans—especially rural Rwandans—are still
very poor, however. More than one-third of the population is
unable to meet its minimum food requirements and rou-
tinely goes hungry (Rwanda Ministry of Finance and
Economic Planning 2007). Although the U.K. overseas devel-
opment agency, the Department for International Develop-
ment (DFID), reports that the poverty rate dropped from
70 percent of the population in 1994 to less than 57 percent
in 2006 (DFID 2008), the World Bank (2010) estimates that
Rwanda’s gross national income per capita was only $460 in
2009. Thus Rwanda has a long way to go to meet the Vision
2020 goal of transforming itself from a low-income to a
middle-income society.
Modest gains and encouraging signs
Changes in the coffee sector began shortly after the geno-
cide, when the government opened the market for coffee
export to increased competition and began to focus on
improving the value chain for coffee (figure 11.1). The
method for pricing beans also changed. Rather than dictat-
ing a single price for the entire season, the Rwanda Coffee
Development Authority (the former OCIR-Café) sets a
minimum weekly reference price, in consultation with
stakeholders, a basis from which a sales price per kilogram
may be negotiated (Mutandwa et al. 2009).5
Figure 11.1 Value Chain for Washed Coffee
Buyers
Farmers private coffee- Millers/
washing stations exporters
or co-op
Source: Author.
188 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
More substantial reform efforts began in 2000, when the
government, working with consultants (the OTF Group has
been the leader in these efforts) and donors, studied the
potential to add value to Rwandan coffee through the pro-
duction of higher-quality, washed, and fermented specialty
coffee. Part of the catalyst for this redoubled effort came as
prices for coffee of ordinary quality fell to a historic low in
1997 and farmers became increasingly unwilling to invest in
the crop. The quantity of coffee produced declined as the
tree stock aged and was not properly tended, soil fertility
declined, and farmers battled with insects and fungal dis-
eases. As a result, the quality of Rwanda’s coffee crop tum-
bled: by 2000, 90 percent of Rwanda’s crop was classified as
low-quality, “ordinary” coffee (Rwanda Ministry of Agricul-
ture and Animal Husbandry and Ministry of Trade and
Industry 2008).
Rwanda’s national coffee strategy
In 2002 the government issued a National Coffee Strategy
that outlined a plan for capturing a larger share of the spe-
cialty coffee sector. A production target of 44,000 tons of
coffee was set for 2010, 63 percent of which would be fully
washed (figure 11.2).6 These production targets have, how-
ever, never been met.
The shift away from low-quality coffee to high-quality
specialty coffee was designed to break a perceived “low-
quality, low-quantity trap.” As noted, Rwandan farmers
were producing low-quality coffee that sold for a low price.
Low sales prices meant that farmers lacked revenue to invest
in improvements. Without sufficient income, farmers could
not invest in capital to improve the quality of their beans—
hence, the trap.
Auction Importers Retailers Consumers
(Mombasa)
Figure 11.2 Targets Set Out in Rwanda’s Fully Washed Coffee Strategy 2002

50,000

45,000

40,000

35,000

30,000 63%
Tons

25,000

20,000

15,000

10,000 25%

5,000
12%
0
02

03

04

05

06

07

08

09

10
20

20

20

20

20

20

20

20

20
Fully washed Standard Ordinary

Source: Rwanda Ministry of Agriculture and Animal Husbandry and Ministry of Trade and Industry (2008).

The 2002 National Coffee Strategy was designed to break
this trap, increase income and revenue, and improve price
stability, because specialty coffee prices fluctuate less dra-
matically than does the price for semiwashed coffee. In an
attempt to meet these targets, coffee-sector stakeholders
focused on strengthening and supporting producer cooper-
atives, identifying sites for and supporting the building of
coffee-washing stations, replanting aging tree stock,
improving quality control throughout the industry, and
strengthening the Rwandan brand globally. And as one of
four focal sectors identified in its 2008–12 Economic Devel-
opment and Poverty Reduction Strategy, the coffee sector
remains a high priority for the government.
established cafés that feature local coffee. In other words,
along a value chain, Rwandans are benefiting from the
opportunity to produce and sell specialty coffee.
At the production level, thousands of Rwanda’s small-
holder farmers are benefiting from higher coffee prices for
fully washed specialty coffee. NGOs such as PEARL and
SPREAD have helped farmers establish cooperatives and
have trained co-op members in quality control, processing,
and marketing efforts. To date, more than 100 washing sta-
tions have been built around the country with the support
of the government, donors, NGOs, and the private sector
(table 11.1). As a result, Rwanda is producing more high-
quality coffee and demand for the country’s specialty coffee
is increasing.

DIRECT ECONOMIC BENEFITS

OF LIBERALIZATION
The liberalization of the coffee sector has had a number of
positive effects. First, farmers now have an incentive—
increased income—to shift some production from semi-
washed to fully washed coffee. Entrepreneurs are investing
in building coffee-washing stations, where cherries are
processed for sale. Rural communities are forming coopera-
tives, some of which also build washing stations and process
cherries. Exporters are competing for opportunities to sell
Rwandan coffee to foreign buyers. Other entrepreneurs have
Incomes are rising for farmers and cooperatives
Perhaps the most important effect of the liberalization of
Rwanda’s coffee industry is that more of the farmers
(approximately 500,000) who grow coffee have an oppor-
tunity to sell their beans for higher prices. The price that
cooperatives and private sector coffee-washing stations are
paying farmers for cherries rose from between 60 and 80
Rwandan francs in 2004 to between 160 and 180 Rwandan
francs in 2008 (Rwanda Ministry of Agriculture and Animal
Husbandry and Ministry of Trade and Industry 2008).7 One

CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS 189
 Table 11.1 Growth in the Specialty Coffee Sector
2002 2003 2004 2005 2006 2007 2008 2009
Number of washing stations 1 10 25 45 76 112 112 112
Tons of green specialty coffee exported 30 300 800 1,200 3,000 2,300 2,455 3,045
Number of specialty coffee buyers 2 8 16 25 30 30 — —
Total value of specialty coffee exported ($1000s) 90 720 1,850 3,168 8,000 7,800 8,060 11,600

Source: USAID (2009).

Note: — Not available.

study finds that farmers who sell coffee cherries to wash-
ing stations increase their annual expenditures by 17 per-
cent compared with farmers who sell lower-quality parch-
ment coffee (Murekezi and Loveridge 2009). The same
study indicates that since reform, coffee farmers have
increased their food consumption and their overall house-
hold expenditures, leading to improved food security and
to generally improved economic conditions for coffee
farmers.
In a 2008 survey of 239 farmers and coffee-washing sta-
tion workers, Tobias and Boudreaux (2011) asked farmers
to identify benefits they received as a result of being a
member of a coffee cooperative.8 Farmers listed a number
of direct financial benefits, such as increased prices
received for their cherries, employment opportunities, and
better and easier access to loans, particularly access to
credit to purchase inputs such as fertilizer. Farmers also
noted that their families are now better fed, that they are
able to hire laborers, that they have help with marketing
and sales, and that they receive some medicines for free.
Some farmers also report that they benefit from access to
coffee bicycles to transport coffee cherries. Less directly,
farmers stated that they benefit from socializing with and
learning from others. Some farmers felt their work was
now easier (they no longer process cherries at home) and
that they could spend time on things other than coffee
production.
A 2006 analysis for the U.S. Agency for International
Development (USAID) reports that “approximately 50,000
households have seen their incomes from coffee production
double, and some 2,000 jobs have been created at coffee
washing stations” (USAID/Chemonics 2006). An NGO
involved in the USAID project (ACDI/VOCA n.d.)
reports that “incomes (in the specialty coffee sector) have
doubled or tripled, and business skills, labour conditions
and community spirit have been enhanced.” With addi-
tional income, farmers can repair their homes, buy clothes,
pay school fees for their children, and get through the long
months between harvests more easily than before (personal
interviews of the COOPAC cooperative 2006).
Once coffee cherries are washed, fermented, and dried,
they can be sold to buyers for an even higher price. In 2004
the Maraba cooperative sold washed coffee for $3.26 a kilo-
gram; in 2007 the cooperative was able to charge $4.08 for
the same amount. The COOPAC cooperative was selling its
cherries for $4.00 a kilogram in 2007, and the Rusenyi coop-
erative was selling in a range between $4.40 and $5.50
(Swanson and Bagaza 2008). On average, in 2007 coffee-
washing stations sold their fully washed coffee for $3.60 a
kilogram. In a notable achievement, importers paid a record
high of $55 a kilogram (approximately $25 per pound) for
the best Rwandan coffee in September 2007, a price compa-
rable to the world’s most expensive coffees (Rwanda Devel-
opment Gateway 2007).
Cooperatives use the income they generate to pay indi-
vidual farmers for cherries, to repay loans taken build wash-
ing stations (or for other equipment), to pay salaries for
washing-station staff, and sometimes to provide other bene-
fits to members—short-term microloans and improvements
to local schools, for example.
Government revenue is increasing
The Rwandan government reports that coffee receipts
increased by an average of 30 percent a year between 2002
and 2006 (figure 11.3). In February 2010 the head of
Rwanda’s Coffee Development Agency said that he expected
26,000 tons of coffee to be produced during the year (up
from 16,000 tons in 2009) and that the value of coffee
exports would increase significantly, to $69 million, from
$38 million in 2009.9
Despite the increases in production and export value, the
amount of fully washed coffee being produced in Rwanda is
still below targets. It now accounts for 20 percent of the
annual crop, compared with 1 percent in 2002. A greater
concern is that washing stations and cooperatives need to

190 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
Figure 11.3 Average Farmer and Export Prices, 2003–08
$3.0
$/kg green coffee
$2.1
$2.0 $1.9
$1.6 $1.8
$1.0
$0.0
2003 2004 2005 2006
Average export price
Cherry price OCIR cafe
Source: Rwanda Ministry of Agriculture and Animal Husbandry and Min-
istry of Trade and Industry (2008).
do more improve technical capacities and operate prof-
itably, creating incentives for more farmers to choose to sell
cherries for washing rather than processing cherries them-
selves at home.
Identifying precisely how much coffee is being produced
in Rwanda is difficult. Government figures from different
sources do not necessarily coincide, and figures from the
U.S. Department of Agriculture and the Food and Agricul-
ture Organization (FAO) do not match figures produced by
the Rwandan government. The government’s trade statistics
put the volume of coffee traded in 2008 at 18,185 tons,
although a more recent news story puts that figure at 22,000
tons.10 The news story indicates that output in 2009 was
approximately 24,000 tons and that the Rwandan govern-
ment projects that output would rise 13 percent in 2010,
to 27,000 tons. Estimates for fully washed coffee are more
solid: in 2002 Rwanda produced 48 tons of fully washed
coffee, a figure that grew exponentially in by 2006, to
3,000 tons (Rwanda Ministry of Finance and Economic
Planning 2007).
Rwanda’s former Minister of Agriculture Anastase
Murekezi has indicated that the specialty coffee industry’s
most successful trading partners to date, in terms of creat-
ing wealth for Rwandans, were American importers. Amer-
ican companies such as Starbucks and Green Mountain
Coffee buy from Rwanda’s producers, bringing much-
needed income to smallholder farmers. Other less-well-known
but highly discriminating American importers such as
Intelligentsia, Thanksgiving, and Counter Culture Coffee
also routinely purchase Rwandan coffee. Coffee buyers
CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
from Europe, China, and Japan also routinely visit Rwanda,
bringing income to farmers and expertise that helps
$3.1 improve the local industry. The benefits from specialty cof-
fee extend beyond the cooperative. Another observer notes
$2.4
that “[a]s income levels of the cooperative members have
increased so has the flow of money in the community . . . .
The positive feelings among community members are a
reflection of increased incomes in the area (of the coopera-
tives)” (Goff 2006, 70).
While global growth in ordinary-grade coffee consump-
tion remains modest, the consumption of high-quality spe-
2007 2008 cialty coffee (currently, 7 percent of the coffee volume in the
Cherry price premium international market) has been rising by 15 percent annu-
ally in recent years. Even with increased competition and
somewhat lower prices, Rwanda’s specialty coffees should
continue to command a good price (Economist Intelligence
Unit 2007).11
Increased entrepreneurship is leading
to more jobs in the coffee sector
As of 2006, 4,000 jobs had been created at coffee-washing
stations (USAID/Chemonics 2006). Though many of these
are part-time jobs during the harvest season, others are
full-time positions managing stations and cooperatives.
Co-op and washing-station employees are learning valuable
business skills: accounting, marketing, and negotiating.
An estimated 100 Rwandans have learned to cup coffee.12
At milling operations and at exporters, other Rwandans sort
beans, operate milling equipment, and prepare beans for
shipment. And as the new coffee culture grows in Rwanda,
jobs are starting to be created in retail outlets such as the
popular Bourbon Café in Kigali.
Although Rwanda has made good progress in the spe-
cialty coffee sector, stakeholders recognize that more must
be done to consolidate these gains and to direct attention
to the most pressing problems in the sector. To refocus
attention, the Rwandan government issued a revised
National Coffee Strategy (NCSR) in 2009. The NCSR cre-
ated a modified strategy for the coffee sector, one that
builds on lessons learned implementing the 2002 strategy.
It sets a new, lower production target of 33,000 tons of
coffee by 2012, 19,000 tons of which is to be fully washed.
The government anticipates revenues of $115 million if
this much coffee is produced (Rwanda Ministry of Agri-
culture and Animal Husbandry and Ministry of Trade and
Industry 2008).
The new policy identifies five target projects: improving
farming techniques, providing support to make washing
191
 stations more profitable, helping private exporters improve
marketing and sales, conducting a census of all coffee-growing
areas, and partnering to do toll roasting in China and the
Middle East.13 In an effort to add more value locally, the
Rwanda Coffee Development Authority (RCDA) has part-
nered with the Hunter Foundation to build a factory to roast
and package coffee locally. The hope is that this effort will
produce more than 100 full-time jobs and another 2,000
indirect jobs.14 While coffee production cannot, by itself,
solve the many problems faced by the poor in Rwanda,
changes in the industry are helping coffee farmers and other
workers in the industry better cope with poverty, and these
changes also seem to be generating positive social benefits.
INDIRECT AND SOCIAL BENEFITS
OF LIBERALIZATION
The liberalization of Rwanda’s coffee sector is distinguish-
able from other liberalization efforts, such as privatizations
in Eastern Europe and land titling reforms in Kenya, that
have tended to benefit elites. Because coffee sector liberal-
ization in Rwanda has helped to raise income for the rural
poor—rather than imposing costs on them—it may be less
likely to promote conflict than have liberalizations in which
costs are spread widely (such as when subsidies are
removed) and benefits are narrowly concentrated (Collier
et al. 2003).15
Coffee facilities are a locus of cooperation
Cooperatives and private washing stations may also serve as
vehicles for increasing cooperation among members. Build-
ing on reports in the media of informal reconciliation at
coffee cooperatives, Tobias and Boudreaux (2011) con-
ducted an exploratory survey of 10 coffee-washing stations
in 2008 to investigate possible social benefits associated with
the liberalization of the sector.16 A total of 239 completed
surveys were obtained from a subsection of rural Rwandans
associated with these stations—some were farmers with sea-
sonal jobs at the washing stations and others were people
who were at the stations to sell cherries. Statistical analysis
of the surveys showed significant correlation between eco-
nomic satisfaction or general perceptions of life satisfaction
and more positive attitudes to reconciliation. This finding
was particularly true the older a washing station was and the
longer it had been a part of the local community.
The survey drew on insights from inter-group contact
theory (Allport 1954). Extensive evidence indicates that
positive interactions between antagonistic groups can
192 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
reduce levels of prejudice and hostility. Therefore, positive
contact is considered one of the most effective strategies for
reducing intergroup conflict (Pettigrew 1998; Dovidio,
Gaertner, and Kawakami 2003). When contact between
groups in postconflict societies is intense and deep, meaning
when it involves interactions at social events, helping others,
or celebrating together, it can promote reconciliation and
help prevent a renewal of violence (Gibson 2004; Straub
2006). If formerly antagonistic groups find ways to cooper-
ate, they may develop a new shared identity that helps
reduce prejudice and creates a sense that a more collabora-
tive future is possible (Gaertner et al. 1990).
As noted, farmers come to coffee-washing stations either
to sell beans or to do seasonal work. Seasonal employees,
working side by side, help at drying tables. During the har-
vest season, farmers may bring cherries to a washing station
several times a week. This means that farmers and seasonal
workers are at the washing stations repeatedly from March
or April through June or July. This organizational structure
is relatively new in Rwanda (in 2000, there were only two
coffee-washing stations in all of Rwanda, and as of 2009
there were more than 100), and farmers may well be experi-
encing a new type of contact.
Perceptions of economic satisfaction
The Tobias and Boudreaux survey explored the possible
connection between perceptions of economic improve-
ments in the lives of farmers and coffee-washing station
workers and of improved interpersonal relations, compar-
ing current levels of economic satisfaction with levels of sat-
isfaction in the past. Economic satisfaction was measured on
a four-item scale, with low scores indicating high levels of sat-
isfaction. A high “economic satisfaction change” score indi-
cated an improvement in economic satisfaction in recent
years. Only 3 percent of participants indicated that they
were very satisfied with their economic situation five years
earlier, whereas 40 percent reported that they are very satis-
fied with their current economic situation. Forty-five
percent of all participants reported a one-point improve-
ment (on a four-item scale) in economic satisfaction in
recent years, 22 percent reported a two-point increase,
10 percent reported a three-point increase in economic
satisfaction, 15 percent experienced no change in economic
satisfaction, and as noted, 3 percent of survey participants
were more satisfied economically in the past. These respon-
dents indicated a decrease of one or two points.
Life satisfaction ratings today and in the past were also
measured. Eighty percent of participants reported a positive
life satisfaction change; for 10 percent, life satisfaction
remained unchanged compared with recent years. Only
7 percent indicated less life satisfaction today than in the
past. These figures, like the figures for economic satisfac-
tion, indicate that the overwhelming majority of the sample
experienced positive life satisfaction gains in recent years.
To explore the impact of increased happiness among
farmers and washing-station workers on other people, the
survey explored possible links between perceptions of eco-
nomic and life satisfaction and willingness to engage in con-
tact with people of other ethnic groups. The findings indi-
cate that people who are less willing to engage socially with
members of another ethnic group have a greater degree of
“ethnic distance” than do people who engage with others
willingly and more frequently. Measures of ethnic distance
were determined in the following manner: an “ethnic dis-
tance today” score was obtained by counting each of five
possible interaction types from a classic social distance scale
(high scores indicate low ethnic distance), and the “ethnic
distance change” score was obtained by calculating the dif-
ference between a positive answer today versus a statement
that in the past the participant would not have engaged in
these interactions. A high numeric score for “ethnic distance
change” was interpreted as a signal of less ethnic distance
today than previously. In general, participants reported high
degrees of reduction in ethnic distance as well as very fre-
quent social and work-related contact.
Survey results show that meaningful contacts with mem-
bers of another ethnic group are significantly correlated
with low distrust and conditional forgiveness. Participants
who expressed satisfaction with their economic and overall
life situation had significantly correlated responses in terms
of positive attitudes toward reconciliation. In particular,
participants with greater economic security also reported
low ethnic distance, low distrust toward the other group, and
a tendency toward conditional forgiveness. Life satisfaction
significantly correlated with economic security variables,
and participants reporting greater satisfaction with life also
expected a more positive, peaceful future in Rwanda.
Results of the survey also show that responses of partici-
pants at coffee-washing stations that had been in operation
for longer periods of time are significantly correlated with a
reduction in ethnic distance over time. It is reasonable to
assume that positive social change in the coffee sector takes
time, and the survey data support this reasoning. All of the
washing stations in the study had been in operation for less
than seven years, and most of them were created fewer than
five years ago. If the observed pattern were to continue,
however, the potential for positive social change associated
CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
with the creation of more than 100 washing stations since
2000 is substantial.
Taken together, the study’s findings suggest that the
enhanced entrepreneurial activities in this particular sector
of Rwanda’s economy not only produce positive economic
change among individuals touched by this institutional
change, but that they may also be triggering a chain of
mediating effects linked to positive social change among
people working at or with coffee-washing stations. The
observed effects were not dependent on ethnicity or on the
particular ethnic mix of participants in a given location.
This finding suggests that forgiveness and increased levels of
trust may be experienced broadly in this environment. Par-
ticipants felt less ethnic distance than in the past and are
now engaging in deeper social contact. It is possible that the
collaboration initiated by the liberalization of the coffee sec-
tor, while difficult to quantify, is one of the most important
benefits of the government’s coffee sector policy reform.
CHALLENGES AND CONCERNS
IN THE COFFEE SECTOR
Over the past decade, the Rwandan government has
worked together with donor organizations, NGOs, the pri-
vate sector (foreign and domestic), and local producers in a
sustained and concerted manner to help this important
sector of the economy become more competitive and, in
turn, grow and expand. Political champions for these
efforts existed at the highest levels of government. Donors
not only supported financing and credit programs, they
also facilitated programs and projects that brought individ-
uals with a variety of technical expertise to Rwanda to share
experiences and knowledge in areas of agronomy, the
development of market links, coffee production, and cup-
ping. Coffee cooperatives received support to improve their
production efforts and to become competitive producers.
Other donor projects supported the entry of Rwandan
entrepreneurs into the market.
These joint efforts required commitments of time,
resources, and talent to address weaknesses in this supply
chain. This iterative process lasted for years. The manner in
which the Rwandan government was able to create an effec-
tive alliance of partners to transform Rwanda’s coffee sector,
through these iterative efforts, from a low-volume, low-
quality market to a high-quality market provides a useful
example of effective policy leadership.
Although the specialty coffee industry is having positive
effects in Rwanda, the smallholder farmers and the entre-
preneurs who work in the sector face several challenges.
193
 One of these is implementing strategies that create price
incentives for farmers to improve both quality and washing-
station management, thereby increasing production effec-
tiveness. At the cooperative level, management needs to be
improved so that costs are better controlled and members
continue to benefit. The government also needs to avoid dis-
locations to poor Rwandans – particularly women – that
may result from the 2005 Land Law. Finally, transportation
costs remain high in Rwanda.
Production capacity and pricing problems
One continuing concern in the coffee sector is that farmers
are not producing a sufficiently large volume of high-
quality coffee to meet continuing demand. A number of
factors may be contributing to this problem. First, prices
paid to farmers do not provide sufficient incentives for
them to focus on quality. Farmers continue to produce a
much greater amount of ordinary-quality coffee than fully
washed coffee. Second, farmers may lack knowledge of how
best to manage their trees. Third, despite the growing num-
ber of washing stations, many farmers still have limited
access to them and so process cherries at home.
As noted above, OCIR sets weekly minimum prices for
the purchase of cherries at washing stations. In some areas,
washing stations pay farmers amounts that fluctuate
modestly to reflect changes in supply. Washing stations do
not, however, seem to pay farmers a premium based on the
quality of cherries they deliver. Farmers thus concentrate on
quantity produced versus quality.
If there is continuing unmet demand for Rwanda’s
fully washed specialty coffee, washing-station operators
should respond by demanding, and paying a premium for,
higher-quality cherries. One possibility why this is not
happening may be that washing-station personnel are
unable to identify higher-quality beans. If washing sta-
tions cannot distinguish a low-quality from a high-quality
cherry, they will choose to pay a low rather than a high
price. Additionally, high labor and input costs also make
it difficult for washing stations to operate at full capacity
and as a result they generate limited income. Therefore,
washing stations may lack sufficient working capital to
pay farmers for higher-quality beans. Competition among
washing stations could resolve this problem. At the
moment, however, it is unclear how best to shift this pat-
tern if there are no other legal impediments to paying
higher prices for higher-quality beans. Although the
problem of pricing is identified in the National Coffee
Strategy Rwanda report, the causes of resistance are not
194 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
identified, suggesting that additional research is needed to
clarify this important issue.
The National Coffee Strategy Rwanda report also notes
that too few farmers are familiar with and able to imple-
ment good farming practices. Farmers need good-quality
inputs, such as seedlings and fertilizers, to increase their
crop yields and quality, but they also need technical knowl-
edge about how to handle seedlings, how to deal with pests
and disease, and how to apply fertilizers. Although
RCDA/OCIR-Café supplies farmers with fertilizer and
employs agronomists, there are only a small number of
these professionals, and they have a difficult time reaching
all the farmers who need assistance.17 The National Coffee
Strategy Rwanda also cites a lack of coordination among
agencies that further limits the effectiveness of these
extension services (Ministry of Agriculture and Animal
Husbandry and Ministry of Trade and Industry 2008).
Management and profitability in Rwanda’s coffee-
washing stations is a growing concern. Stations are operat-
ing at partial capacity, producing smaller amounts of fully
washed coffee than anticipated. The government reports
that operating costs associated with the stations (labor,
transportation, and electricity and water) are high. In
addition to these challenges, washing-station personnel
often lack technical skills, particularly in finance, account-
ing, and management. These capacity problems can trans-
late into difficulties securing financing and working with
lenders, which in turn can translate into difficulties paying
farmers in a timely fashion. The government recognizes
the potential benefits of increasing coffee yields at washing
stations, particularly for fully washed coffee. Farmers who
sell coffee as a cash crop are projected to see their incomes
grow 4 percent a year more than that of other farmers
(Ministry of Agriculture and Animal Husbandry and
Ministry of Trade and Industry 2008).
Concerns with cooperative management
Smallholder farmers who voluntarily join cooperatives
face the difficult task of creating a culture of entrepre-
neurship within the cooperatives so that they become
more “business minded.” A key problem identified by
local NGO SPREAD is the need to attract and retain more
professional managers in cooperatives and, at the same
time, to reduce the influence of volunteer boards of direc-
tors. An assessment of a group of Rwandan cooperatives
states that “[a] professional, entrepreneurial [g]eneral
[m]anager is the most important individual to a cooper-
ative’s ultimate success” (Swanson and Bagaza 2008).
However, the report’s authors find that no cooperatives in
the group under investigation had such a manager. The rea-
son seems to be that the boards of directors are reluctant to
pay high enough salaries to attract a professional manager.
Further, boards often prefer to have to have a local repre-
sentative, rather than an “outsider,” fill the general manager
role. Local representatives, however, are less likely to have
the skill set needed to manage the cooperative effectively.
Cooperatives are capable of producing very high-quality
coffee, but they are experiencing real difficulties creating
effective management structures.
Rwanda’s coffee producers, with support from donors,
NGOs, and the government, have done an impressive job of
generating interest in their products. Rwandan coffee is reg-
ularly available at retailers such as Starbucks and Whole
Foods across the United States and Marks & Spencer in the
United Kingdom. But this good progress may stall if prob-
lems related to marketing and processing are not adequately
addressed. Buyers have expressed concerns that contract
terms for quality, quantity, and timely delivery are not being
met (Rwanda Ministry of Agriculture and Animal Hus-
bandry and Ministry of Trade and Industry 2008). Some
buyers have reported long delays receiving shipments, and
many have complained that the shipped product did not
match samples. These kinds of concerns, if not addressed,
will lead to loss of business for Rwanda.
A related problem is that boards of directors in Rwanda
often interfere inappropriately in the daily management
of the cooperative. The assessment notes that “[board
members], particularly [p]residents, do not want to relin-
quish their authority to a strong [g]eneral [m]anager”
(Rwanda Ministry of Agriculture and Animal Husbandry
and Ministry of Trade and Industry 2008, 13). While it is
essential that boards take seriously their fiduciary duties to
create general policies and oversee management activities,
the assessment recommends that they give managers
increased decision-making authority and discretion.
Managers of cooperatives and boards of directors need to
communicate more effectively with members so that mem-
bers understand the ownership structure of the cooperative
as well as their rights within the structure. To date, this has
not been done effectively. Members report that they are
unclear who “owns” the cooperative (Rwanda Ministry of
Agriculture and Animal Husbandry and Ministry of Trade
and Industry 2008). In addition, cooperatives need to
develop and communicate effective business plans and to
improve financial record keeping and documentation.
There is a clear need for increased capacity in these areas.
When members lack clear information about the financial
CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
state of the cooperative, and about likely prices for cherries
and benefits of membership, the possibility of corruption
and conflict over resources rises. These are especially impor-
tant issues to resolve because cooperatives face increasing
competition from other coffee entrepreneurs.
Cooperatives have been an important asset for
Rwanda’s smallholder farmers, allowing them to earn
more money from coffee, develop additional skills, and
work cooperatively with others in ways that may promote
reconciliation. But they must address serious shortcom-
ings in management practices and capabilities if they hope
to continue playing this role in the future. Because coop-
eratives seem to provide a space for interaction and even
informal reconciliation, further support efforts to help
accomplish the goal of creating transparent and account-
able management may well be justified.
Concerns with Rwanda’s land law
A separate challenge in Rwanda involves use and control of
land. Land is an extremely scarce and highly contested
resource. More than three-fifths of families working as
farmers cultivate less than 0.7 hectares, and more than a
quarter cultivate less than 0.25 hectares (Rwanda Ministry
of Finance and Economic Planning 2007). According to one
study, “land was a factor behind social tensions before every
major open conflict. Even today, more than 80 percent of all
disputes in Rwanda are related to land” (Food and Agriculture
Organization and Rwanda Ministry of Lands, Environment
Forestry, Water and Mines 2006, 7; see also Van Hoyweghen
1999). For much of its history, Rwanda’s rulers have owned
most of the country’s land. With control of land in the
hands of government, formal land markets did not develop.
Rather, transfers often took place informally, and confusion
and insecurity were common. Local officials had great dis-
cretion over land allocation and favored politically power-
ful individuals over marginalized people who may have held
traditional land use rights.18
In 2003 the Rwandan parliament approved a land reform
decree that provides for individualized rights to property.
This policy was followed, in 2005, by passage of the Organic
Land Law. The law is implemented by a series of decrees,
many of which are just now going into effect. The Land Law
abolishes all customary forms of tenure, and in their place,
establishes government-issued titles for 99-year leases. Rural
land will be registered locally, and urban, commercial prop-
erty will be registered in a national cadastre in Kigali. The
government maintains a role in the resettling of people and
in devising land use and land planning policy.
195
 The government “sees increased security of tenure or
rights of address to land, and more effective land manage-
ment, as important factors for the improvement of the agri-
cultural sector and the economy as a whole, helping to create
the resources needed to reduce poverty and to consolidate
peace and social cohesion” (Pottier 2005, 511). Although
the government says that it wants to increase tenure secu-
rity, the new law is likely to create a host of problems.
The government hopes the land law will promote con-
solidation of land parcels into larger units. By allowing sales
of property and increasing freedom within the land market,
small parcels could be sold to commercial farmers who will
consolidate the land and create viable agribusinesses. How-
ever, one observer notes:
Land fragmentation in Rwanda serv[es] as a coping
mechanism in smallholder agriculture, the typical
Rwandan household farms an average of five plots.
Some are in the valleys, others are upland and some
near the household. In some parts of southern
Rwanda, a household may have up to 14 crops grow-
ing in different fragments at different seasons. . . . the
costs of consolidation in Rwanda may not exceed the
benefits of using land fragmented over the years in
adopting to land scarcity (Musahara 2006, 11).
While a more open land market is desirable, the law
(Organic Law No. 08.2005) contains provisions that allow
the government to interfere in the market. For example, it
allows the government to bar people who own less than one
hectare from registering their property. More troubling is
the provision that “subsistence farmers can have their land
confiscated should they fail to exploit it diligently and effi-
ciently” (Pottier 2005, 521, [Articles 62–65]). Though the
government is supposed to provide compensation for such
confiscations, it has not established clear standards for such
payments.
The new law may pose a special problem for women
smallholders and their children. Under the law, the gov-
ernment is supposed to register all parcels of land in the
country. Only legally married women and their children,
however, (not women married under customary norms and
their children, or poor women who do not formally marry
because of the associated costs) can register and inherit
land. Further, there is uncertainty in the law regarding
inheritance—specifically, whether women inherit through
the inheritance law or through the land law.19 Also of con-
cern is the fact that local custom continues to bar women
from exercising their legal rights under the land law.
196 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
Security and clarity of tenure rights, whether customary
or leasehold, are essential both to avoid future conflicts and
to encourage increased investment in agriculture. However,
the land law raises serious concerns, especially for women
and for uneducated farmers who might be dispossessed of
their land. Surely, these risks are undesirable in a nation
with such high levels of poverty and such strong depen-
dence on agriculture as a livelihood.
Concerns with transportation infrastructure
Finally, as a landlocked country with limited paved roads in
rural areas where most coffee is grown, transport costs in
Rwanda are high. Diop, Brenton, and Asarkaya (2005) argue
that Rwanda’s smallholder subsistence farmers are discon-
nected from markets as a result of “extremely high” transport
costs. The authors estimate that transport costs from farm
gates to the export port of Mombasa were 80 percent of the
producer price.20 Transport within Rwanda itself was esti-
mated at 40 percent of the producer price. If transport costs
were reduced, through the development of better rural infra-
structure and, in particular, more effective rural transport
routes, access to markets would improve and poverty levels
would likely be reduced. The authors find that a 50 percent
reduction in the transport costs in rural areas would lead to a
20 percent increase in producer prices for coffee, which in turn
would reduce poverty levels among coffee farmers by more
than 6 percent. Given the continued emphasis on coffee pro-
duction as a strategy to alleviate rural poverty, improving the
rural transport system will be an important way to connect
farmers to markets and to increase their household income.
CONCLUSION
Despite good economic growth and real benefits of liberal-
izing the coffee sector, more remains to be done to move
Rwanda toward its Vision 2020 goal of becoming a stable,
middle-income country. Most Rwandans remain very poor
and the rural areas of the country provide few off-farm
employment opportunities. Creating more employment
opportunities for millions of Rwandan smallholder farmers
remains a pressing challenge.
Rwanda’s specialty coffee industry is helping to address
some of these concerns. As a result of improvements in the
sector, smallholder coffee farmers now keep more of the
value of the product they grow. In turn, rising incomes for
tens of thousands of farmers make them better able to feed
themselves and their families, to send their children to
school, to buy insurance, and to repair or improve their
homes. By freeing the coffee sector from the heavy-handed
involvement of the government, the Rwandan government
has shifted incentives in the coffee sector and created greater
scope for citizens to pursue entrepreneurial opportunities.
Though additional research needs to be done, evidence
suggests that joint efforts in commercial activities related to
the specialty coffee sector are providing an alternative path
to postwar reconciliation in Rwanda. Initial exploratory
survey data provide some supports for these claims (Tobias
and Boudreaux, 2011). By working together in cooperatives
and at washing stations, Rwandans are experiencing both
social and economic benefits from liberalization. They feel
greater economic satisfaction, which is correlated to lower
levels of distrust, higher levels of forgiveness, and reduced
ethnic distance.
Whether this kind of reconciliation is effective in the
medium to long term in Rwanda remains to be seen. The
positive experience of workers in the coffee sector, however,
strongly suggests that in postconflict environments, govern-
ments should follow Rwanda’s lead and promote broad-based
trade liberalization that encourages commercial interaction
between rival groups.
Although real concerns surround the economic viability
of some coffee-washing stations and cooperatives, liberal-
ization of the Rwandan coffee sector has created a wider and
deeper space for positive entrepreneurship: a space being
filled by thousands of Rwandans, from smallholder farmers
to local exporters. At the same time, adding value to the cof-
fee supply chain is producing both direct economic benefits
and important indirect social benefits for individuals and
communities in Rwanda.
NOTES
1. “Coffee ‘Cup of Excellence’ Slated for Next Year,” New
Times, September 23, 2009.
2. Traditionally, Rwandan farmers removed the fruit of
their coffee cherries using either a hand-pulper or rocks.
Beans would then be dried and fermented in buckets, for
varying lengths of time, in water of varying quality. As a
result, coffee was of lower, industrial quality. This home-
processed coffee still makes up the majority of coffee being
sold from Rwanda.
3. Prices were set by OCIR until 1994 and remained con-
stant for the entire harvest season; price differences based
on quality were not permitted. See Mutandwa et al. 2009
and deLucco 2006 for more information.
4. Under the Habyarimana regime, the powerful OCIR-
Café agency was run by relatives of the dictator’s wife, mem-
bers of the clan de Madame (Verwimp 2003).
CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
5. Mutandwa et al. also note that “due to the low level of
production, the milling factories operate under capacity and
exporters tend to lower the reference price in order to cover
their relatively high milling costs” (p. 212).
6. A conflicting statement, however, can be found in the
Economic Development and Poverty Reduction Strategy
2008–2012, which states that the “the proportion of fully-
washed coffee production will increase from 10 percent to
100 percent . . .” (p. 38).
7. See also Swanson and Bagaza (2008).
8. Survey results are available from the author of this paper.
9. Reuters, “Rwanda Sees 20010 Coffee Crop Rising 63 Pct,”
February 11, 2010.
10. “Coffee Output May Climb 13 Percent in 2010,” New
Times, January 8, 2010.
11. The report notes, “Stagnation in many of the tradi-
tional coffee-drinking markets of North America and West-
ern Europe will restrict growth in demand, although
demand for high-quality specialty coffees, including
Rwanda’s finest fully-washed Arabica, will remain more
buoyant” ( p. 10).
12. To “cup” coffee refers to a tasting technique used to cre-
ate a flavor profile for coffee. Cuppers evaluate coffee based
on fragrance, aroma, and taste, a process that helps identify
defects in coffee. Fewer than half of the 100 cuppers, how-
ever, are trained and licensed to international standards.
“Twenty-One Coffee Cuppers Trained on Quality,” New
Times, August 29, 2009.
13. Toll roasting involves a contractual arrangement
whereby an experienced coffee roaster accepts green coffee
from a client and roasts it to the client’s specifications.
14. “Two Million Invested in Coffee Value Addition,” New
Times, August 7, 2009.
15. Collier et al. (2003, 53) argue that “the key root cause of
conflict is the failure of economic development. Countries
with low, stagnant, and unequally distributed per capita
incomes that have remained dependent on primary com-
modities for their exports face dangerously high risks of
prolonged conflict.”
16. Surveys were conducted at five cooperatives and five
privately owned coffee-washing stations. Surveys were
administered in Kinyarwanda by students from the National
University of Rwanda over a two-week period. Surveys were
previously field tested in February 2008.
17. Farmers receive fertilizer from RCDA, and the agency, in
turn, imposes a fertilizer fee on exporters. Exporters then
choose either to pass these costs along to cooperatives and
washing stations in the form of lower prices paid per kilo or
they absorb the costs themselves.
18. Sales were restricted according the size of the buyer and
seller’s total land holdings; see Musahara (2006).
197
 19. Dyer (2007) notes that although Rwanda’s 1999 Inheri-
tance Law gives women equal rights to men this has “yet to
make any difference in practice.”
20. Note these prices should now be lower as a result of
Rwanda’s accession to the East African Community.
REFERENCES
ACDI/VOCA. n.d. “Specialty Coffee: Increased Quality &
Profits for Smallholders.” http://www.acdivoca.org/
852571DC00681414/Lookup/coffeebroweblayout/$file/
coffeebroweblayout.pdf.
Allport, Gordon W. 1954. The Nature of Prejudice. Oxford,
UK: Addison Wesley.
Bates, Robert H. 1981. Markets and States in Tropical Africa:
The Political Basis of Agricultural Policies. Berkeley, CA:
University of California Press.
Berlage, Lode, Bart Capéau, and Philip Verwimp. 2004.
“Dictatorship in a Single Export Crop Economy.” House-
holds in Conflict Network. http://www.hicn.org/papers/
dicta.pdf.
Coffeeresearch.org. “Coffee Cupping.” http://www.coffeere-
search.org/coffee/cupping.htm.
Collier, Paul, V. L. Elliott, Håvard Hegre, Anke Hoeffler,
Marta Reynal-Querol, and Nicholas Sambanis. 2003.
Breaking the Conflict Trap: Civil War and Development
Policy. Oxford, UK: Oxford University Press and World
Bank.
deLucco, Paul. 2006. “Raising the Bar: Producing Quality
Coffee in Rwanda.” World Report Spring 2006: Specialty
Coffee; Improved Market Linkages & Increased Profits.
http://www.acdivoca.org/852571DC00681414/Lookup/
WRSpring06-Page13-14-RaisingtheBar/$file/WRS
pring06-Page13-14-RaisingtheBar.pdf.
DFID (UK Department for International Development).
2008. “Good Taxes Reduce Poverty in Rwanda.” January 10.
Diop, Ndiame, Paul Brenton, and Yakup Asarkaya. 2005.
“Trade Costs, Export Development, and Poverty in
Rwanda.” Policy Research Working Paper 3784, World
Bank, Washington, DC.
Dorsey, Learthen. 1983. The Rwandan Colonial Economy:
1916–1941. Unpublished PhD dissertation, Michigan
State University.
Dovidio, John F., Samuel L. Gaertner, and K. Kawakami.
2003. “Intergroup Contact: The Past, the Present and
the Future.” Group Processes and Intergroup Relations
6: 5–21.
Dyer, Rodney. 2007. “Land Issues in Rwanda: Drawn from
DFID Technical Assistance for Land Tenure Reform,
2005–2008.” http://www.oxfam.org.uk/search?q=dyer&
x=17&y=19.
198 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
Economist Intelligence Unit. 2007. “Country Report:
Rwanda” (February).
FAO (Food and Agriculture Organization) and Rwanda
Ministry of Lands, Environment Forestry, Water and
Mines. 2006. “A Case Study on the Implications of the
Ongoing Land Reform on Sustainable Rural Develop-
ment and Poverty Reduction in Rwanda and the Out-
come Report of the Thematic Dialogue Held on the 20th
January 2006.” Presented at the International Conference
on Agrarian Reform and Rural Development (ICAARD),
Porto Alegre, Brazil, March 7–10. http://www.icarrd.org/
en/icard_doc_down/case_Rwanda.pdf.
Gaertner, Samuel L., Jeffrey A. Mann, John L. Dovidio,
Audrey J. Murrell, and Marina Pomare. 1990. “How Does
Cooperation Reduce Intergroup Bias?” Journal of Person-
ality and Social Psychology 59 (4): 692–704.
Gibson, James L. 2004. “Does Truth Lead to Reconciliation?
Testing the Causal Assumptions of the South African
Truth and Reconciliation Commission.” American Jour-
nal of Political Science 48 (2): 201–17.
Goff, Samuel Neal. 2006. “A Case Study of the Management
of Cooperatives in Rwanda.” Unpublished master’s the-
sis, Texas A&M University.
IFAD (International Fund for Agricultural Development).
2006. “Enabling the Rural Poor to Overcome Poverty
in Rwanda.” http://www.ifad.org/operations/projects/
regions/Pf/factsheets/rwanda_e.pdf.
Index Mundi. 2010. “Rwanda GDP – Real Growth Rate.”
http://www.indexmundi.com/rwanda/gdp_real_growth_
rate.html.
Murekezi, Abdoul, and Scott Loveridge. 2009. “Have Coffee
Reforms and Coffee Supply Chains Affected Farmers’
Income? The Case of Coffee Growers in Rwanda.” Paper
presented at the Agricultural & Applied Economics
Association’s 2009 Joint Annual Meeting, Milwaukee,
WI, July 26–28.
Musahara, Herman. 2006. “Improving Tenure Security for
the Rural Poor: Rwanda—Country Case Study.”
FAO/LEP Working Paper 7.
Mutandwa, Edward, Nathan Taremwa Kanuma, Emmanuel
Rusatira, Theopile Kwiringirimana, Patrice Mugenzi,
Ignatius Govere, and Richard Foti. 2009. “Analysis of
Coffee Export Marketing in Rwanda: Application of the
Boston Consulting Group Matrix.” African Journal of
Business Management 3 (5): 210–19.
Personal interview with Anastase Murekezi, Rwandan
Minister of Agriculture. 2006 (March). Kigali, Rwanda.
Personal interviews with members of the COOPAC cooper-
ative. 2006 (March). Gisenyi, Rwanda.
Pettigrew, T. F. 1998. “Intergroup Contact: Theory, Research
and New Perspectives.” Annual Review of Psychology 49:
65–85.
Pottier, Johan. 2005. “Land Reform for Peace? Rwanda’s
2005 Land Law in Context.” Journal of Agrarian Change
6(4): 509–37.
Prunier, Gerard. 1995. The Rwanda Crisis: History of a
Genocide. New York: Columbia University Press.
Republic of Rwanda. Organic Law No. 08.2005: Determin-
ing the Use and Management of Land in Rwanda, Article
20. Kigali. July 14.
Rwanda Development Gateway. 2007. “Coffee Sells at
Record Prices.” http://www.rwandagateway.org/article
.php3?id_article=6848.
Rwanda Ministry of Agriculture and Animal Husbandry
and Ministry of Trade and Industry. 2008. “National Cof-
fee Strategy Rwanda 2009–2012.”http://amis.minagri
.gov.rw/sites/default/files/user/National_Coffee_
Strategy_Rwanda_2009-2012.pdf.
Rwanda Ministry of Finance and Economic Planning. 2000.
“Rwanda Vision 2020.” http://www.gesci.org/assets/files/
Rwanda_Vision_2020.pdf.
———. 2007. “Economic Development and Poverty
Reduction Strategy 2008-2012.” http://planipolis
.iiep.unesco.org/upload/Rwanda/Rwanda_EDPRS_
2008-2012.pdf.
SPREAD. 2007. Annual Report. http://www.spread.org.rw/
spread_project.php.
CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS
Straub, Edwin. 2006. “Reconciliation after Genocide, Mass
Killing or Intractable Conflict: Understanding the Roots
of Violence, Psychological Recovery and Steps toward a
General Theory.” Political Psychology 27 (6): 867–95.
Swanson, Richard, and Tom Bagaza. 2008. “SPREAD-Growers
First Coffee Cooperative Assessment and 2008 Coopera-
tive Development Work Plan.” Report prepared for the
Texas A&M University System, Borlaug Institute of Inter-
national Agriculture and Growers First.
Tobias, Jutta, and Karol Boudreaux. 2011. “Entrepreneur-
ship and Conflict Reducation in the Post-genocide
Rwanda Coffee Industry.” Journal of Small Business and
Entrepreneurship 24 (2): 217–242.
USAID (U.S. Agency for International Development). 2009.
“Rwanda.” http://www.usaid.gov/rw/our_work/programs/
docs/factsheets/coffee.pdf.
USAID/Chemonics. 2006. “Assessing USAID’s Investments
in Rwanda’s Coffee Sector: Best Practices and Lesson
Learned to Consolidate Results and Expand Impact.”
Van Hoyweghen, Saskia. 1999. “The Urgency of Land and
Agrarian Reform in Rwanda.” African Affairs 98: 353–72.
Verwimp, Philip. 2003. “The Political Economy of Coffee.”
European Journal of Political Economy 19 (2): 161–81.
World Bank. 2010. “Country Data: Rwanda.” http://data
.worldbank.org/country/rwanda.
199
 CHAPTER 12
Cocoa in Ghana: Shaping the Success
of an Economy
Shashi Kolavalli and Marcella Vigneri
o other country comes to mind more than Ghana
N
when one speaks of cocoa. Likewise, one cannot
think of Ghana without thinking of its cocoa
sector, which offers livelihoods for over 700,000 farmers in
the southern tropical belt of the country. Long one of
Ghana’s main exports, cocoa has been central to the coun-
try’s debates on development, reforms, and poverty allevia-
tion strategies since independence in 1957. The cocoa sector
in Ghana has not been an unmitigated success, however.
After emerging as one of the world’s leading producers of
cocoa, Ghana experienced a major decline in production in
the 1960s and 1970s, and the sector nearly collapsed in the
early 1980s. Production steadily recovered in the mid-1980s
after the introduction of economywide reforms, and the
1990s marked the beginning of a revival, with production
nearly doubling between 2001 and 2003. These ups and
downs offer interesting lessons.
Various administrations in Ghana, including the colonial
one, have used cocoa as a source of public revenue, and in
so doing the Ghanaian experience offers a recurrent exam-
ple of a policy practice followed by many other African
countries: taxing the country’s major export sector to
finance public expenditure (Herbst 1993). Revenue extrac-
tion by the state has had varying effects on production
Shashi Kolavalli is senior research fellow and leader, Ghana Strategy Support Program, International Food Policy Research Institute. Marcella Vigneri is a
visiting scholar, Department of Economics, University of Oxford.
depending on global prices, marketing costs, explicit taxes
on the sector, and macroeconomic conditions such as infla-
tion and overvaluation of exchange rates and inelasticity of
cocoa supplies. Regardless of the level of extraction, the
need for sound macroeconomic management, of inflation
and exchange rates in particular, becomes evident for con-
tinuing to offer incentives for production. The other is the
need for Ghana’s cocoa pricing policy to arrive at a market-
ing arrangement that does not kill the goose that lays the
golden eggs. Ghana appears to have achieved such as
arrangement without fully liberalizing the sector as other
producers in West Africa have.
OBSERVABLE ACHIEVEMENTS
IN THE COCOA SECTOR
Since the introduction of cocoa in Ghana in the late 19th
century, the crop has undergone a series of major expansions
and contractions. Ruf and Siswoputranto (1995) suggest that
cycles are intrinsic to cocoa production because cocoa is
influenced by environmental factors such as availability of
forest land; ecological factors such as deforestation, out-
breaks of disease, and geographic shifts in production; and
economic and social factors such as migration.
201
 Emergence as a leading producer buying companies on the coast of West Africa that were pre-
pared to trade the new crop (Hill 1963; Amanor 2010; and
Four distinct phases can be identified in regard to cocoa
Gunnarsson 1978).
production in Ghana: introduction and exponential growth
Three social classes: land-owning farmers, peasants, and
(1888–1937); stagnation followed by a brief but rapid
laborers emerged among cocoa producers as a second wave of
growth following the country’s independence (1938–64);
migrants from Akyem moved to the region. Without sufficient
near collapse (1965–82); and recovery and expansion, start-
money with which to buy land, these migrants sharecropped
ing with the introduction of the Economic Recovery Pro-
with earlier settlers under a system called abusa, in which
gram (ERP) (1983 to present). Figure 12.1 shows long-term
laborers were paid one-third of the sales price of the harvested
trends in levels of production.
cocoa. Simultaneously, there was a large influx of migrants
from relatively distant Upper Volta (now Burkina Faso), Niger,
Exponential growth (1888–1937). Cocoa was introduced
and Mali, who were attracted by the generous remuneration
in the southern region of the Gold Coast in the mid-19th
that cocoa production offered in southern Ghana.
century by commercial farmers from the Eastern region dis-
The growing population of cocoa farmers reinvested its
tricts of Akuapem and Krobo, who had moved west toward
profits in cocoa production in the western end of Ghana’s
the adjacent district of Akyem to purchase mostly unoccu-
Forest Zone, rapidly shifting the production frontier into
pied forest land from the local chiefs for cocoa cultivation
the Ashanti and Brong Ahafo regions, and consolidating
(Hill 1963).
Ghana as the leading world producer between 1910 and
The conditions that encouraged these farmers to migrate
1914. Facilitated by the rapid expansion of the road and rail
and buy land for cocoa are well documented: a fall in the
network which began in 1920 and the organization of cocoa
world price of palm oil after 1885, which pushed farmers to
marketing by Ghanaian middlemen, cocoa earnings
search for alternative export crops; a boom in rubber
accounted for 84 percent of the country’s total exports by
exports in 1890, which provided the capital for the purchase
1927. By the mid-1930s, production reached 300,000 tons.
of new land; increasing population pressure in the
Akuapem area, which encouraged commercial farmers to
go further afield in search of alternative export agriculture Stagnation and growth postindependence (1938–early
opportunities; and the establishment of European produce- 1964). The interwar period marked a slowdown in cocoa

Figure 12.1 Ghana’s Cocoa Production, 1900–2008

800

700

600
Production, thousands of tons

500

400

300

200

100

0
19 0
19 3
06
19 9
12
19 5
19 8
19 1
24
19 7
30
19 3
19 6
39
19 2
19 5
48
19 1
54
19 7
19 0
63
19 6
69
19 2
19 5
78
19 1
84
19 7
90
19 3
96
20 9
02
20 5
08
0
0

1
1
2

3
3

4
4

5
6

7
7

0
19

19

19

19

19

19

19

19

19

19

19

19

19

19

20

Source: Gill & Duffus Group, various issues; Ghana Cocoa Marketing Board, various issues.

202 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY

production, caused by decreasing demand and growing dif-
ficulties in transport (Gunnarsson 1978). Outbreaks of
pests and diseases (swollen shoot virus in particular)
reduced production in the Eastern region in the early 1940s,
pushing cocoa cultivation further into the western Brong
Ahafo frontier (Amanor 2010). Production picked up again
during the second half of the 1940s but was now concen-
trated in the Western region. In 1947, the colonial govern-
ment established the Cocoa Marketing Board (CMB) and
gave it a monopoly over the purchase of beans. Until 1951
the bulk of profit made by the CMB went into its reserves,
which were then used for public investment (Brooks, Crop-
penstedt, and Aggrey-Fynn 2007). In 1961 a cooperative
society was given the monopoly right to purchase cocoa
replacing the network of private agents, brokers, traders,
and middlemen who until then had controlled internal
marketing.
As Beckam (1976) noted, the Convention People’s Party
(CPP), founded by Kwame Nkrumah, benefited from
extremely favorable postwar market conditions and accu-
mulated cocoa income on a massive scale: following the
sharp increase in market prices in the 1950s, farmers were
paid two to three times more than they received before the
war, and between 1947 and 1965 the government collected
almost one-third of the total value of cocoa export as export
duties. In 1950/51 the government increased export duties
and began to take a much larger share of cocoa revenue by
means of a graduated ad valorem tax that increased with the
increase of the average selling price per ton of cocoa. To
extend its influence to the rural sector, in 1953 the Nkrumah
regime also created the United Ghana Farmers’ Council
(UGFCC), which was mainly concentrated in the cocoa-
growing regions despite its remit to cover the interest of
farmers all over the country. The UGFCC was made the
monopoly buyer of cocoa to create a platform for organizing
the farmers behind the government and its administration.
Following the second elections in 1954, the cocoa export
tax was further increased while the producer price remained
at the same level for four years. This generated unrest and
political agitation among cocoa farmers, ultimately forcing
the government to increase the producer prices and to sta-
bilize them during 1956–57 despite declining world cocoa
prices. As a result, the share of government revenue in cocoa
sales dropped from 60 percent to 13 percent between
1954/55 and 1956/57. After its third political victory of
1957, the government increased its share of cocoa revenues
by reducing producer prices to the 1954 levels. It also
obtained a “voluntary contribution,” announced by the
UGFCC on behalf of cocoa farmers, to share the burden of
CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
the Second Development Plan at a time when the govern-
ment was also receiving soft loans from the CMB. These
events made it obvious that by then the CMB had been
transformed into an instrument of public finance. The cap-
turing of windfall profits from high cocoa prices had impor-
tant fiscal implications. Government expenditures grew
dramatically over the 1950s: in real terms total consolidated
public expenditures increased almost sixfold during this
period. The share of government expenditure in GDP grew
from 7 percent to 18 percent over the decade, and the share
of extraordinary and development expenditure grew from
27 percent to 36 percent. In 1961, a cooperative society was
given the monopoly right to purchase cocoa. From 1957 to
1964 exports grew steadily, and production reached an
unprecedented level of 430,000 tons despite the significant
decline in world prices between 1960 and 1962.
In the early 1960s, when world prices plummeted, farm-
ers were required to save 10 percent of their earnings in
National Development Bonds, redeemable after 10 years. In
1963 this scheme was replaced by a farmers’ income tax
charged at a flat rate equal to previous saving deductions.
The government started to rely heavily on the CMB’s
reserves, and the producer price was reduced from 224 to
187 new cedi per ton between 1961 and 1964. With foreign
exchange reserves declining and the budget deficit rising
sharply, the government introduced a number of strong
restrictive measures, an increase in taxes, foreign exchange
controls, and comprehensive import licensing. The austerity
of these measures lost Nkrumah much of his political con-
sent, especially from cocoa farmers who had been aggra-
vated by declining producer prices and by the conversion of
the compulsory saving scheme into an explicit export tax.
In the second half of 1964 the world cocoa price collapsed
with a bumper crop in West Africa—Ghana alone reaching
an unprecedented production record of 538,000 tons. After
the purchasing and marketing costs of the CMB and UGFCC
were covered, virtually nothing was left for the government,
and the CMB’s liquidity resources were nearly exhausted. To
meet its expenses, the government started printing money,
which ignited a 35 percent rise in inflation between October
1964 and July 1965. In the face of such pressure, cocoa pro-
ducer prices were reduced to their lowest levels in years. The
introduction of such highly restrictive measures represented
a turning point in the fortunes of the Nkrumah government,
which was overthrown in February 1966 and replaced by the
National Liberation Council (NLC).
The downturn (1964–82). The collapse of world cocoa
prices in 1965 triggered another downturn (Stryker 1990).
203
 Real producer prices dropped consistently through the
1960s because of inflation fueled by the government’s print-
ing of money to compensate for loss of revenue from cocoa
and the introduction of an exchange rate policy that led to
the heavy overvaluation of the cedi, the local currency. By
1983, market exchange rates were nearly 44 times the official
rate. Between 1970s and early 1980s, it is estimated that as
much as 20 percent of Ghana’s cocoa harvest was smuggled
into Côte d’Ivoire (Bulír 2002). Meanwhile, an aging tree
stock and the continued spread of disease made investment
in cocoa unattractive. Farmers in old cocoa production
areas, who found that sales prices barely covered their costs,
increasingly turned from cocoa to food production
(Amanor 2005). Ghana’s cocoa production dipped to a low
of 159,000 tons in 1982/83, a mere 17 percent of the total
world volume, down from the 36 percent in 1964/65.
The National Liberation Council dissolved the UGFCC
and established the Producing Buying Company as a sub-
sidiary of the CMB. Producer prices were raised and farm-
ers were paid a bonus for top grade cocoa beans to upgrade
the quality of cocoa being exported. Shortly before the
Busia government came to power the cedi was devalued by
43 percent and cocoa prices were raised by 30 percent.
Cocoa production stagnated in the face of unchanged real
producer prices that remained at their 1950s levels. The
Busia administration took advantage of windfall profits
from high cocoa prices in 1970 to enable a rapid expansion
of public expenditure.
In 1971 the Busia regime was replaced by the
Acheampong-led National Redemption Council. Because of
high world cocoa prices, this administration was initially
able to offer higher prices to farmers without cutting public
revenues, creating positive incentives to production. But a
progressively worsening balance of payments situation
fueled inflation and undermined subsequent increases in
real wages, producer prices, and other real incentives.
With the fall in world cocoa prices in the mid 70’s, the
general macroeconomic picture began to worsen: the gov-
ernment budget deficit rose to 127 percent of total govern-
ment revenue and inflation accelerated to 116 percent. The
strong overvaluation of the cedi implied that little was left
of export revenues to divide between the government and
the farmers. Cocoa revenue went from 46 percent in 1974
to 23 percent in 1979 and into negative figures between
1980 and 1981 because of the exchange rate misalignment.
The rising costs of the CMB further reduced government
revenues.
In July 1978 the government underwent another regime
change, and the cedi was devalued again, an austerity budget
204 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
was introduced, and interest rates and cocoa producer
prices were raised. Cocoa production sunk to its lowest level
ever in 1980–81; the world price at the official exchange rate
was lower than the producer price plus marketing costs.
The domestic conditions that led to the downturn in
Ghana’s cocoa sector took place against an international
backdrop of increasing supply of cocoa from new producers
such as Indonesia and Malaysia and expanded production
in Côte d’Ivoire and Brazil. By the early 1970s Ghana had
also lost much of its cheap labor supply from Burkina Faso
and Côte d’Ivoire, as migrant farmers, reluctant to work in
the old cocoa-producing areas that had become less pro-
ductive, were attracted to the neighbouring Ivorian
regions, where policies granted migrants access to land at
favorable terms.
The recovery and second expansion phase (1983–2008).
The turnaround in Ghana’s cocoa sector began with the
implementation of the ERP in 1983, which included a spe-
cial program to revive the sector (the Cocoa Rehabilitation
Project). Policy changes included increasing the farm gate
prices paid to Ghanaian farmers relative to those paid in
neighboring countries, thus minimizing the incentive to
smuggle, and devaluing the cedi, thus reducing the level of
implicit taxation of farmers.
As part of the Cocoa Rehabilitation Project, farmers
were also compensated for removing trees infected with
swollen shoot virus and planting new ones. This effort led
to substantial rehabilitation, with a large number of farms
planting higher-yielding cocoa tree varieties developed by
the Cocoa Research Institute of Ghana. Production
rebounded to 400,000 tons by 1995/96 and productivity
increased from 210 to 404 kilograms per hectare. Another
important reform took place in 1992, when Cocobod (as
CMB was renamed in 1984) shifted responsibility for
domestic cocoa procurement to six privately licensed com-
panies (commonly known as licensed buying companies
or LBCs) and reduced its staff by 90 percent between 1992
and 1995.
Growth in cocoa production became more pronounced
starting in 2001, possibly driven by a combination of
record-high world prices, increased share being passed onto
farmers, and a set of interventions rolled out by the Coco-
bod to improve farming practices: mass spraying programs
and high-tech subsidy packages to promote the adoption of
higher and more frequent applications of fertilizer (Vigneri
and Santos 2008). Some of the growth during this period
may also have been due to the influx of cocoa smuggled
from Côte d’Ivoire. One study estimated this amount at
between 120,000 and 150,000 thousand tons in 2003/4
Table 12.1 Fertilizer Use in Cocoa-Producing
(Brooks, Croppenstedt, and Aggrey-Fynn 2007). Regions: 1991/92–2003/04
Brong
Technical change in the cocoa sector Crop year Ashanti Ahafo Western Total
Number of farmers
Since 2001 a significant share of Ghana’s agricultural produc- 1991/92 112 71 137 320
tivity gains have been generated by export crops, with cocoa 1997/98 132 54 227 413
accounting for 10 percent of total crop and livestock produc- 2001/02 108 94 226 428
2003/04 108 94 226 428
tion values (World Bank 2007a) and contributing to 28 per- Quantity of fertilizer used (50-kilogram bags)
cent of agricultural growth in 2006, up from 19 percent in 1990/91 0.28 0.13 0.03 0.14
2001. At the same time, economic growth has been solid, Adoption rate (%) (13) (8) (6) (9)
1997/98 0.10 0.06 0.10 0.09
averaging more than 5 percent since 2001 and reaching 6 per- Adoption rate (%) (10) (13) (19) (15)
cent in 2005–06. Coupled with the effects of greater access to 2001/02 0.35 0.17 0.74 0.52
education, health services, and land ownership (World Bank Adoption rate (%) (5) (7) (12) (9)
2003/04 4.17 4.39 6.10 5.24
2008), this rate of growth has contributed to the near halving
Adoption rate (%) (57) (52) (41) (47)
of the national poverty rate since the beginning of the 1990s,
from 51.7 percent in 1991/92 to 28.5 percent in 2005/06 Source: Authors’ calculations from GLSS3, GLSS4, and Ghana Cocoa
Farmers Survey, 2002 and 2004 rounds.
(Breisinger et al. 2008).
Over time, cocoa farmers have changed the way they
access land and labor in response to the changing produc- markets in 1996/97, which eliminated subsidies but
tion conditions of a constantly moving cocoa frontier. Until improved private distribution (Vigneri and Teal 2004).
the early 1940s, when both land and labor were abundant,
large farms were able to attract rural workers to establish Adoption of improved varieties. Hybrid cocoa varieties
new farms by selling them small plots of land, an arrange- were introduced in 1984 through the government’s Cocoa
ment that often also drew the workers’ family members to Rehabilitation Project (CRP). Hybrid varieties outperform
establish and maintain new farms. By the second half of the the older “Amazons” and “Amelonado” varieties in two
1960s, when land became scarce, sharecropping arrange- ways—by producing trees that bear fruit in three years com-
ments increasingly replaced land sales. During times when pared with at least five years for the older varieties, and by
the cost of hiring waged workers became too high, alterna- producing more pods per tree.1 But hybrid cocoa trees
tive forms of labor were used—mostly, either sharecrop- underperform older varieties in that they require optimal
ping arrangements or informal labor groups known as weather conditions and complementary farming practices
nnoboa (Berry 1993; Blowfield 1993; Vigneri, Teal, and such as the application of chemical inputs, adoption of new
Maamah 2004; and Amanor 2010). Since 1990 noticeable planting procedures, pruning, and spraying. Hybrids vari-
changes have taken place in the technology of cocoa pro- eties also require that farmers make more harvest rounds at
duction, in particular increased use of fertilizers; the adop- the beginning and the end of the season, something they are
tion of hybrid cocoa varieties, and better control of pests reluctant to do when it conflicts with other farming or trad-
and diseased trees (Boahene, Snijders, and Folmer 1999; ing activities (Boahene, Snijders, and Folmer 1999; Bloom-
Edwin and Masters 2003; Gockowski and Sonwa 2007; Teal, field and Lass 1992).
Zeitlin, and Maamah 2006; Vigneri, Teal, and Maamah Despite the increased labor input for hybrid cocoa trees,
2004; and Vigneri 2008). farmers have increasingly adopted them. In the late 1980s
only 10 percent of cocoa grown in Ghana was of the high-
Increased use of fertilizer. Fertilizer use in Ghana has yielding type (Nyanteng 1993). By 2002, 57 percent of farm-
increased significantly since the 1990s. Surveys of cocoa ers in the three main cocoa-producing areas were growing
farmers in the three main cocoa-producing regions of hybrid trees (Vigneri 2005). Traditional varieties may have
Ghana show that fertilizer application rates increased from disappeared entirely from all fields planted after 1995
9 percent in 1991 to 47 percent in 2003 (table 12.1). (Edwin and Masters 2003).
Although the quantity of fertilizer used decreased between
1991/92 and 1997/98, the proportion of farmers applying Better disease and pest control. Control of disease
fertilizer increased, possibly from liberalization of input and pests, swollen shoot virus and capsid in particular,

CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 205

 has improved significantly in recent years. After Cocobod
initiated a free mass spraying program in 2001, 93 percent
of cocoa farmers who participated in a survey conducted
in 2002 linked their yield improvements to the effects of
the program (Steedman 2003). Similarly the cocoa farm-
ers’ panel referred to above for the crop years 2002 and
2004 suggests that nearly all farms were sprayed in
2003/04, when producers reported an average of more
than four spraying applications during the crop year, of
which 46 percent were carried out by the government
(Vigneri 2005).
The effect of all these improved practices has been an
increase in productivity of about 30 percent, which brought
productivity to the levels achieved in the 1980s (figure 12.2).
Productivity was stagnant until the late 1980s, with produc-
tion largely related to area harvested. The first big jump in
productivity occurred in the 1980s, corresponding to the
year of the Cocoa Rehabilitation Program rolled out under
the ERP, and the second more recently, with improved prac-
tices. The correlation between production and area har-
vested remains strong.
Cocoa’s contribution to economic growth and
poverty reduction
In the Southern Forest Belt, where cocoa is produced, aggre-
gate figures suggest that through the 1990s, cocoa-farming
households, along with those engaged in mining or timber
(the other predominantly export-oriented activities) and
other commercial activities, experienced improvements in
their living conditions compared with food crop farmers
(McKay and Coulombe 2003). Poverty reduction among
cocoa farmers is clear. Household surveys indicate that
poverty among cocoa-producing households dropped to
23.9 percent in 2005, down from 60.1 percent at the begin-
ning of the 1990s (World Bank 2007b).
Reputation for high-quality cocoa
Cocoa, like many other commodities, is often differentiated
by country of origin, and this in turn is associated with a
reputation based on average quality. The reputation, a
national public good, enables the country to earn a pre-
mium in the global market for the crop it is producing.
Generally, Ghana receives a price premium for its cocoa in
world markets because of the slightly higher-than-average
fat content; low levels of debris, which results in higher
cocoa butter yields than beans containing high levels of
debris; and low levels of bean defects, which generate a
cocoa liquor flavor preferred by some end users. In addi-
tion to these attributes, the reputation of the Cocoa Mar-
keting Company (the government division in charge of all
exports) in ensuring the consistency and reliability of
cocoa-related shipments and documents has played a cen-
tral role in establishing the country’s reputation for high-
quality beans (Agrisystems Ltd. 1997). Using trade NYSE
Liffe cocoa market information, Gilbert (2009) suggests
that Ghanaian cocoa draws a premium of 3 to 5 percent rel-
ative to Côte d’Ivoire, currently the world’s largest producer
of cocoa (table 12.2).

Figure 12.2 Cocoa Production, Area Cultivated, and Yields, 1961–2008

2.5 800

700
Millions of hectares harvested

2.0
600
Yields (kgs/ha)

1.5 500

400
1.0 300

200
0.5
100

0 0
61

19 3
65
67

19 9
71

19 3
75
77

19 9
19 1
19 3
85
87
89
91
93

19 5
19 7
99
01

20 3
05
07
6

7
8
8

9
9

0
19
19

19
19

19

19
19

19
19
19
19
19

20
20

20

Area harvested, left hand side Production (000 tons), right hand side
Yields (Kgs/ha), right hand side

Source: FAOSTAT.

206 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY

 Table 12.2 Cocoa Unit Values and Terminal Market Differentials
Percent
Cameroon Ghana Nigeria
Period Unit value Differential Unit value Differential Unit value Differential
1988–1991 2.7 — 3.7 — –0.4 —
1992–2002 –3.0 0.20 1.1 4.8 –2.1 –0.5
2003–2008 –7.8 — 5.2 4.9 –0.7 –0.9
1988–2008 –3.3 — 2.8 4.9 –1.4 –0.7

Source: Adapted from Gilbert (2009). Figures reported are relative to those of Côte d’Ivoire, the reference country.
Note: — = not available.

Characteristics that determine the quality of cocoa Figure 12.3 Share of Processed Cocoa Products in Total
include content and quality of fat, consistency in the size of Cocoa Exports in West African Countries, 1990–2007
the beans, and their moisture content. These characteristics
30
determine the quality of cocoa butter and cocoa liquor
produced from the beans, the two ingredients that control 25
texture, aroma, color, and flavor of chocolate. The fermen-
20
tation, drying, storage, and evacuation of wet beans can
Percent
alter the quality of cocoa beans dramatically, particularly in 15
the development of the flavor of cocoa liquor. The classic 10
“West African” cocoa flavor is obtained by fermenting beans
in a heap under banana leaves for about six days with fre- 5

quent manual turning and thorough drying in the sun. 0

Drying beans slowly on raised platforms is very important
90

92

94

96

98

00

02

04

06
19

19

19

19

19

20

20

20

for the quality of flavor because it quickly decreases the
acidity level of the beans. Quality is also maintained by
quickly collecting properly fermented and dried beans from
smallholder farmers and promptly shipping them to avoid
the buildup of moisture, mold, and free-fatty acids that can
rapidly deteriorate the quality of the bean.
Partly because of its reputation for high-quality cocoa,
Ghana is able to sell most of its annual production
through forward contracts, which fix the price farmers are
given for their cocoa for the entire crop year. The value
that international firms place on Ghana’s cocoa is also
reflected by the amount of investment they have made in
processing facilities in the country. Ghana’s export earn-
ings from processed cocoa products more than tripled
between 1991 and 2004, from $32 million to $105 million
(figure 12.3).2 However, because of the limited conditions
under which semiprocessed cocoa can be transported effec-
tively (Fold 2002), it is not clear whether local value-adding
efforts will be sufficiently profitable for international com-
panies to expand their operations in Ghana. Thus far,
informal discussions with the private sector participants
indicate that the net benefits from processing locally may
not be significant, particularly because the government
allows only a limited quantity of low-quality beans to be
20
Cameroon Côte d'Ivoire Ghana Nigeria
Source: FAOSTAT.
Note: Cocoa processed products include cocoa butter and cocoa paste.
used for local processing, which has resulted in consider-
able underutilization of existing capacity in the country.
Increased share of free on board prices
going to farmers
Agricultural exports continue to be the most important
source of foreign exchange for the majority of Sub-Saharan
African countries (Gilbert 2009). In virtually every country
in Africa with a major export crop, including Ghana, the
government has intervened through state-owned marketing
boards, or caisses de stabilization, to coordinate the produc-
tion and marketing of the crop, offering farmers stable farm
gate price that shield them from price volatility. Many schol-
ars (Bates 2005; McMillan 1998; Akiyama et al. 2001) hold
that marketing boards in Africa have long operated as cor-
rupt institutions taxing farmers through the power to set
prices and indirectly by maintaining overvalued exchange

CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 207

 rates. That said, the role of governments in the agricultural
sector has changed substantially since independence.
Despite granting Cocobod the monopoly over market-
ing, Ghana has managed to develop a marketing system that
passes on an increasingly larger share of export prices to
farmers. Prices received by Ghanaian producers have been a
function of government interest in using the sector as a
source of revenue and a balance against global prices,
exchange rate distortions, and inflation. Price policies were
also made ineffective by macroeconomic policies. In the
early 1980s, for example, the Provisional National Defense
Council (PNDC) had to choose between supporting cocoa
farmers and continuing to maintain highly overvalued
exchange rates (Stryker 1990).
In Ghana, the price producers are paid for cocoa is cur-
rently set at the beginning of the harvest season for the entire
crop year by the Producer Price Committee.3 The price is
based on the price Cocobod expects to receive, having
already sold nearly 70 percent of the crop. To this price,
Cocobod adds the costs of its operations and the export tax
to arrive at what it calls “net free on board (f.o.b.) price.”
The share of the net f.o.b. price received by cocoa farm-
ers in Ghana has increased to nearly 80 percent after having
fallen below 20 percent before the economic reforms of the
1980s, and as low as almost 5 percent between 1975 and
1981. By 1987/88, real producer prices in Ghana had
increased threefold compared with 1983/84, largely as a
result of Cocobod’s revised policy of paying higher prices to
the farmers, in response to pressure from multilateral
organizations to streamline its operations (Brooks, Crop-
penstedt, and Aggrey-Fynn 2007). Figure 12.4 shows the
share of f.o.b. prices paid to producers, the share retained by
Cocobod (shown as direct taxation), and the share of indi-
rect taxation imposed by the exchange rate.4 Exchange rate
distortions can further erode the share producers receive.
These distortions were high in the mid-1980s but have com-
pletely disappeared.
REASONS FOR SUCCESS OF THE
COCOA SECTOR
A number of factors have contributed to the success of
Ghana’s cocoa sector: a favorable price regime, both in
terms of the f.o.b. share passed on to producers and the real
price received by farmers; improved marketing through par-
tial liberalization; and Cocobod’s interventions to raise
cocoa productivity.
Favorable prices
With the exception of 1998–2000 and 2003–06, world cocoa
prices have steadily increased since 1990. This, combined
with a higher share of the price being passed on to farm-
ers, has offered farmers increasing real producer prices
(figure 12.5). A variety of models estimating the sensitiv-
ity of production supply to farm gate prices find that

Figure 12.4 Farm Gate Prices, Direct Taxation, and Exchange Rate Taxation for Ghanaian Cocoa, 1966–2008
Percent

120

100

80

60

40

20

–20
19 67
19 69
19 /71

19 73
19 /75
19 /77
19 79
19 /81
19 83
19 85
19 87
19 /89
19 /91
19 /93
19 /95
19 /97
20 /99
20 /01

20 03
20 /05
20 /07
9
/0
/
/

/
/
/

/
66
68
70
72

74
76
78
80
82
84
86
88
90
92
94
96
98
00
02

04
06
08
19

Farm gate price Direct taxation Exchange rate taxation

Source: Vigneri 2005.

208 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY

Figure 12.5 Ghana Cocoa Production and Real Producer Price, 1990–2008

180,000 800

160,000 700
140,000

Price per ton (2005 cedis)

600
120,000
500

Metric tons
100,000
400
80,000
300
60,000
200
40,000

20,000 100

0 0
19 0
19 1
19 2
93

19 4
19 5
19 6
97

19 8
20 9
20 0
20 1
02

20 3
20 4
05

20 6
07
08
9
9
9

9
9
9

9
9
0
0

0
0

0
19

19

19

20

20

20
Real producer price Production

Source: Cocobod and ICCO.

small-scale cocoa producers in Ghana have responded posi-
tively to these price incentives (Bulíř 2002; Hattink, Heerink,
and Thijssen 1998; and Vigneri 2005, among others).
Although strictly comparable data are not available,
informed inference on the returns on cocoa farms using the
results from two rural surveys, one conducted in 1996
(Agrisystems Ltd. 1997) and one in 2006 (Barrientos and
Asenso-Okyere 2008), show that cocoa production has not
become more profitable for farmers. In fact, calculations
show that cocoa, which usually is the largest source of earn-
ings in cocoa-producing households, accounting for more
than 67 percent of revenues, has actually declined over time:
net cocoa profits for cocoa-producing households were
7 percent lower in 2005 than in 1996. While the real price of
cocoa increased by 47 percent between these two years, the
cost of inputs increased more. These estimates, however, do
not suggest a trend because they are based on observations
for two specific years.
Liberalization of domestic cocoa marketing
Following pressures from multilateral organizations in the
early 1980s, wide-ranging changes were introduced to
improve Cocobod’s efficiency: transport was shifted to the
private sector, feeder road development was transferred to
the Ministry of Roads and Highways, and in 1988–89 input
subsidies were phased out (Brooks, Croppenstedt, and
Aggrey-Fynn 2007). Following the 1992 elections more dras-
tic measures were undertaken: Cocobod staff levels were
reduced from 100,000 in the early 1980s to 10,400 in 1999 to
just over 5,100 in 2003, bringing down costs considerably. In
the same year, Cocobod ended its control over all domestic
purchases by allowing a number of private licensed compa-
nies to compete with its former purchasing agency, the
Producing Buying Company (PBC), to buy and transport
the cocoa crop from farms; the board, however, specifies a
minimum price. This partial liberalization appears to have
benefited producers. The internal marketing of cocoa has
also become more competitive in recent years, with nearly 20
licensed buyers, along with PBC, procuring cocoa through
nearly 3,000 buying stations manned by purchasing clerks or
individuals from cocoa communities who purchase the crop
on the buyers’ behalf. Although the total number of licensed
buyers is relatively large, five dominate the market: the Pro-
duce Buying Company, Kuapa Kokoo, Olam, Armajaro, and
Global Haulage, a former transport company comprising
three Ghanaian buyers (Federal Commodities, Transroyal,
and Adwumapa) Additionally, Cocobod extends funds to
producers at rates slightly below the market rate to finance
their operations. It also monitors producers’ operations, par-
ticularly with regard to quality of beans. Though licensed
buyers are free to export, none of them has thus far because
none is large enough to acquire the minimum amount
needed to be eligible to export.
Zeitlin (2006) finds a positive correlation between the
concentration of licensed buying companies at the village
level and production. But the direction of causality is not
clear, because buyers are also likely to locate themselves

CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 209

 where large quantities of cocoa are available for purchase.
The PBC continues to operate as a buyer of last resort.
While Cocobod sets a minimum price that must be paid to
producers, the buying companies are free to pay higher
prices. Even in the absence of price competition among
licensed buyers, farmers have benefited. Payments to farm-
ers have become more reliable, and corruption, which char-
acterized the contractual negotiations when the PBC was
the only buyer, has diminished. While licensed buyers may
not compete on prices, they do offer occasional price
bonuses, subsidized inputs, or credit extensions for produc-
ers (Laven 2007). Because the licensed buyers buy cocoa
throughout the year, however, the new buying system puts a
steadier stream of money into the hands of producers
(Vigneri and Santos 2008), giving farmers working capital
to buy labor and other inputs when they need them.
Although the efforts at liberalization are likely to have
made procurement and transport more efficient than
before, it is unclear whether Cocobod’s costs have been
reduced by “outsourcing” procurement and transport and
to what extent liberalization may have helped Cocobod pass
on a higher share of f.o.b. prices to farmers. But regardless,
retaining control over exports and other aspects of market-
ing has enabled Cocobod to support producers in ways that
would not have been feasible had it devolved these respon-
sibilities to other organizations.
Cocobod’s impact on productivity
Importantly, Cocobod’s continued involvement in the cocoa
sector in Ghana has allowed surpluses generated in good
years to be used to finance deficits during years when prices
were low. Similarly, Cocobod has invested in research, dis-
ease control, and credit programs that are of general benefit
to the cocoa industry (Stryker 1990). In 2001 the Cocoa
National Disease and Pest Control Committee was estab-
lished to develop strategies to control capsid and black pod
through a nationally coordinated spraying program under
which Cocobod, through a network of regional offices,
undertakes spraying of all cocoa fields at no cost to the
producers.5 By Cocobod’s estimates, the scheme has had a
positive impact on national cocoa production, particularly
during the 2003/04 and 2005/06 seasons. Cocobod also
reports that the protection of the cocoa plants that the pro-
gram offers has encouraged farmers to undertake additional
spraying applications.
In 2002/03, Cocobod rolled out the “Cocoa High-Tech”
program designed to encourage farmers to apply a minimum
of 5 bags of fertilizer per hectare of planted cocoa, supplying
210 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
fertilizer on credit. The program collapsed after one year,
however, because of poor repayment rates. Following this
pilot, a private agri-input company, Wienco, tested a package
of agricultural inputs and farm practices known as the
“Abrabopa package.” In 2003, its first year of testing, the pack-
age raised yields from 510 to 1,081 kilograms per hectare and
to 2,317 kilograms per hectare after the third year.
In 2006 the Cocoa Abrabopa Association (CAA) was
established, under which groups of farmers with mature
trees on at least one hectare of land were given the Abrabopa
package on credit and offered technical and business train-
ing. The number of farmers participating in this program
reached 11,000 in 2008. An evaluation of the program in
2008 (Opoku et al. 2009) suggests that the principle of
group liability employed in this program ensured, to some
extent, the effective use of the fertilizer and other inputs
provided by the CAA package. That said, a large proportion
of farmers, nearly 40 percent, dropped out of the program,
so the benefits of the CAA package reached only a small
share of cocoa growers.
Cocobod’s role in maintaining quality
In terms of quality practices by government marketing
boards among West African cocoa-producing countries,
Ghana is an exception, because maintenance of quality
continues to be Cocobod’s mandate even after its restruc-
turing. In other countries, dismantling and restructuring of
marketing boards in the 1980s radically reduced quality
control systems (Fold 2001; Gilbert 2009). One rationale for
a government role in maintaining quality is that cocoa is
transported in bulk, and poor-quality cocoa beans can
diminish the quality of other beans in the same shipment,
thereby affecting the price of all beans in the shipment.
Maintaining a government role is also important because it
allows the government to control the national reputation of
Ghana’s cocoa and keep its premium in the world market
(Fold and Ponte 2008). This quality maintenance comes at a
cost, however, including the cost of ensuring that lower-
quality beans are not mixed into those prepared for export
and the costs of administration.
SUSTAINABILITY OF THE COCOA SECTOR
Ghana’s cocoa sector faces a number of challenges. For one,
productivity levels are lower than they are in other countries.
Ghana also faces the possibility that its quality advantage
may disappear in the coming years. In addition, Ghana must
determine how to keep its cocoa sector competitive as
cocoa-producing households change. Finally, the environ-
mental impact of current farming practices may soon con-
strain cocoa production expansion. On the other hand,
however, Ghana has been quite successful in taking advan-
tage of niche cocoa markets.
Productivity and competitiveness
Notwithstanding the technical changes that have occurred in
cocoa production, Ghana still needs to close a large produc-
tivity gap to remain competitive. The gap between observed
and achievable yields is 50–80 percent (Gockowski 2007),
depending on the production practices adopted by farmers
(for example, thin shading and the amount of fertilizer
applied). A survey conducted in the 1980s, however, indi-
cated that Ghana was the lowest-cost producer in the world
(Bloomfield and Lass 1992). Ghana’s yields are low com-
pared to those of its leading competitors, Côte d’Ivoire and
Indonesia (figure 12.6). Additionally, it is not clear which
technologies intended to increase productivity are attractive
to farmers. For example, farmers may not have much incen-
tive to apply fertilizers to hybrid trees, because the returns
from doing so may not be higher than those achieved on tra-
ditional varieties (Edwin and Masters 2003).
On experimental farms, application of fertilizers to
young trees has increased yields as much as threefold
Figure 12.6 Cocoa Yields, by Country, 1990–2008
1,200
1,100
1,000
900
kg per hectare
800
700
600
500
400
300
200
1990 1992 1994
Cameroon
Source: FAOSTAT.
(Gockowski and Sonwa 2007). One evaluation (Opoku et al.
2009) suggests that the high dropout rate from the CAA
program may result from high variability in the expected
returns from fertilizer applications.
The low level of tree replanting is an additional threat to
the sustainability of Ghana’s cocoa production. Often, farm-
ers find it more economical to expand their farms rather
than to replace old and diseased trees (Vigneri 2005; Ruf
and Burger 2001), because it takes twice as long to clear an
old farm as it does to clear new forest land (Masdar Ltd.
1998). Additionally, farmers regard the expansion of land on
which cocoa is planted as both an investment and a means
to establish land ownership. Given that migrants and share-
croppers represent an increasing share of the cocoa-farming
population, this dual view means that many farmers seek to
acquire permanent land rights by expanding into unculti-
vated land, where land ownership is established by clearing
land and planting new trees (Amanor 2010; Berry 2009;
Takane 2002).6 Further opportunities to increase produc-
tion by land expansion may be limited, though, by the
decreasing availability of virgin forest land.
Longevity of the quality advantage
Although Ghanaian cocoa draws a premium price for its
reliable quality, this advantage may be eroded in the future
1996 1998 2000 2002 2004 2006 2008
Côte d'Ivoire Ghana Indonesia Nigeria
CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 211
 because of technological advances in processing
(Agrisystems Ltd. 1997; Fold 2001). On the other hand,
current quality control processes in Ghana guarantee mini-
mum parameters that are important to large industry
players like Cadbury, which is known to use Ghanaian cocoa
beans exclusively in all its U.K.-retailed chocolate products.
A second, potential threat is when other cocoa-producing
countries improve on their quality. Currently, this is not
much of a threat to Ghana, because smallholder farmers in
countries such as Malaysia and Indonesia lack the institu-
tions to support quality. In Côte d’Ivoire, the mixing of
good cocoa beans with the bad ones in shipments for export
results in variability in quality (Bloomfield and Lass 1992).
Competitiveness of cocoa on farms
Cocoa is a mixed crop system in which other crops may be
consumed or sold. Intercropping with plantain and
cocoyam, for example, provides early returns when cocoa
trees are still young. Studies conducted in the 1970s and in
the 1990s (Rourke 1974; Masdar 1998) report that almost all
cocoa farmers grew alternative crops for subsistence and
sale, mostly roots and tubers but also a variety of cereals and
vegetables. Both studies also suggest that many farmers
shifted to crops other than cocoa (mixed plantain and
cocoyam, mixed maize and cassava, and oil palm inter-
cropped with maize and cassava) on a scale greater than that
needed to satisfy subsistence needs. This shift occurred for
several reasons: the crops offered farmers greater income
continuity throughout the year, and returns were perceived
to be higher relative to cocoa, especially in the presence of
significant problems with the rehabilitation of the existing
cocoa tree stock.
More recent research has questioned the viability of
cocoa on small farms. A 2001 survey conducted by the Sus-
tainable Tree Crop Programme (STCP) in four cocoa-
producing countries in West Africa shows that the top 25
percent of households (ranked by the amount of cocoa
produced) have average costs of production four times
lower and yields nearly four times greater than the bottom
25 percent, and that a significant share of small cocoa
farms incur losses (Gockowski 2007). The study recom-
mends the urgent adoption of policies that vary for larger,
more efficient producers and poorer marginal ones as a
necessary step in keeping Ghana’s cocoa sector competitive
and efficient.
For the larger producers, the STCP study recommends
implementing innovations through the strategic distribu-
tion of improved planting material (hybrid pods) in the
212 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
most densely populated regions of the cocoa belt. The study
estimates that this could result in the replanting of up to
24,000 hectares of land, and that integrating this interven-
tion with the expansion of fertilizer use would achieve pro-
ductivity gains in excess of 50 percent. For less efficient
cocoa producers, the STCP recommends implementing a
different set of policies that would either allow these produc-
ers to exit the sector or support their transition to alternative
production systems. One option for these less efficient farm-
ers would be the conversion from a no-shade cocoa system to
a partial-shade system with cocoa and non-cocoa trees inter-
cropped, allowing producers to augment their incomes from
the sale of forest products, and possibly from the additional
payments for higher carbon sequestration associated with
shaded tree systems.
Environmental impact of current
farming practices
An issue closely related to the competitiveness of cocoa on
farms is the environmental impact of existing farming prac-
tices. Since its introduction in West Africa, cocoa has been
the major cause of land use change in the high forest zones
of the regions in which it is grown, where it has replaced
agricultural activity that incorporated fallowing to maintain
land fertility (Gockowski and Sonwa 2007). Although the
initial expansion of cocoa production did not entail a com-
plete removal of the forest shade because the traditional
shade-dependent and tolerant tetteh quarshie variety of
cocoa did not require forest clearing, trees have been cut
down en masse in recent years to accommodate the open-
field hybrid variety, which grows in full sun conditions. In
nearly three-quarters of Ghana’s production area, there is
little to no shade (table 12.3).
Farmers in Ghana have a strong preference for full-sun
crops because their much shorter growing cycle is linked to
higher short-term profits (Obiri et al. 2007). The damage to
cocoa trees from capsid attacks tends to be higher for cocoa
trees growing in full sun than for those in shaded systems,
Table 12.3 Shade Levels in the Cocoa Belt of Ghana
(percent)
Region None to light Medium to heavy
Ashanti 52 47
Brong Ahafo 52 47
Eastern 50 49
Western 77 21
Ghana 72 29
Source: Adapted from Gockowski and Sonwa (2007).
however, and the carbon sequestration potential of full-sun
cocoa systems is significantly less than that of traditional
shaded cocoa systems (Norris 2008).
The best possible environmental alternative to the cur-
rent cocoa-growing practices in Ghana would be a mixed
agroforestry system, where the forest is selectively thinned
and fruit trees with economic value—such as oil palm, avo-
cado, and citrus—are grown next to cocoa trees, providing
both shade for the cocoa trees and food and income for the
farming household (Gockowski and Sonwa 2007). This
practice, which is used in southern Cameroon, could offer
farmers up to 23 percent of total revenues from their non-
cocoa holdings, but it is rarely practiced in Ghana. One rea-
son is that in the biodiversity hotspots in remote areas of the
Western region, the profitable marketing of agroforestry
products would not be easy. Additionally, past logging prac-
tices, in which concessionaires harvested in a way that
destroyed cocoa farms with no compensation for producers,
have discouraged the use of fruit and timber-producing
trees in cocoa fields (Obiri et al. 2007).
Ghana’s role in a changing global
market for cocoa
Ghana is well positioned to expand its position in high-
value markets, with Cocobod proving to be responsive to
trends in international markets. The chocolate industry also
has expanded into secondary markets, such as fair trade in
the late 1980s. Although these markets offer strategic oppor-
tunities for countries to build competitiveness, estimated in
2000 at 2.6 percent of world cocoa bean trade (Abbott
2002), they largely remain niche markets because of their
limited capacity for expansion.
Ghana’s considerable progress in the fair-trade cocoa
market began with the establishment, in 1993, of Kuapa
Kokoo, a farmers’ cooperative that operates as a private,
licensed buying company. Its share in the domestic market
is now estimated to be around 10 percent of total purchases,
and a panel survey of farmers spanning 2002 to 2006 shows
the cooperative to be farmers’ second preferred outlet for
selling beans (Vigneri and Santos 2008). Within Cocobod, a
special channel exists for fair-trade cocoa sourced and
exported from Kuapa Kokoo, although the system traces
such cocoa back to the cooperative rather than to the indi-
vidual farmer. The social premium earned on fair-trade
exports, which in 2000 was reported to be $150 per ton
(Abbott 2002), goes into a trust fund that sponsors develop-
ment projects in cocoa-producing communities. Recently,
the CAA became the first cocoa cooperative in Ghana to
CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
obtain certification for organic production, with more than
500 members meeting the required standards.
LESSONS FROM GHANA’S EXPERIENCE
WITH COCOA
Cocoa was developed in Ghana, largely by commercial
farmers, many of whom were smallholders and laborers
drawing on their own savings and labor, in response to mar-
ket opportunities and the development of infrastructure.
Policies and institutions have played an important role. The
importance of macroeconomic management, the avoidance
of distortions in the exchange rate in particular, is clearly
evident from the effect of its absence on farm gate prices in
the mid-1980s.
Ghana appears to have emerged with an appropriate
institutional mix in which competition has been introduced
in internal marketing to benefit from efficiencies in pro-
curement and transport, while the government marketing
board retains control over setting minimum prices for the
year, maintaining quality, and managing exports. The con-
trol it has retained over exports enables it to stabilize prices
and use the surpluses to offer some services such as plant
protection, research, and extension that may not be forth-
coming from the private sector, as suggested by the experi-
ence of the fully liberalized producing countries in the
region. Public support to farmers to rehabilitate the dis-
eased tree stock, public research that produced new hybrids,
and the continued state intervention to promote fertilizer
use have all been instrumental in reviving the sector. More
recently public spraying and dissemination of technical
packages have spurred private action.
Would the cocoa sector have been better if it were fully
liberalized? Examining the experience of liberalization of
cocoa sectors in four West African countries, Gilbert (2009)
suggests four criteria to address this question: (1) the level
of competition achieved on both the export and import side
of producing countries, (2) the ability to sustain quality
standards, (3) the share of the f.o.b. price passed on to the
farmers as an indication of the degree of state taxation, and
(4) the extent of producer price stabilization achieved. The
evidence suggests that it may not be so.
Liberalization has not resulted in competition in the
value chain, particularly in exports. Local companies
engaged in exports without access to global financing have
withdrawn over the years, leaving exports largely in the
hands of multinationals, either converters or their agents
(Gilbert 2009). But, there has been greater competition in
internal trade. As for the share of the f.o.b. price passed on
213
 to the farmers, the proportion is higher in countries such as
Cameroon and Nigeria, but Ghana’s government has made
concrete efforts in the recent past to raise the share similarly.
Finally, in relation to the price stabilization objective, Ghana
has clearly been successful in reducing farmers’ exposure to
price variability during the crop year through its practice of
forward sales. This, combined with the more stable inflation
rate of the past decade, has de facto acted as an insurance
mechanism against the variability in the world price of the
commodity. Global businesses like the Ghana model
because it delivers consistently high quality. Local businesses
are also content because they can continue to participate in
the sector (Gilbert 2009).
The interesting question is whether it is possible to arrive
at this mix of public and private institutions and also be cer-
tain that a parastatal organization such as the Cocobod
would operate reasonably efficiently. Ghana’s experience
suggests that external pressures as a part of the ERP to
reform the sector were instrumental in making Cocobod lib-
eralize some of its operations and streamline its own work-
ing to reduce costs. Ghana appears to have done enough to
fend off pressures for further liberalization of the sector. To
what extent it will strive to continue to pass on a higher share
of prices to farmers without external pressures and whether
there is a recognition of the benefits from appropriate man-
agement that survives political changes are not clear. The
affairs of the Cocobod are not as transparent as they should
be, and the line between cocoa revenues and government
finances remains fuzzy. Whether the Cocobod will be able to
stabilize prices if the world market were to become more
volatile than it has been in recent years is not clear.
The pressure on the government and on its marketing
institution to improve their efficiency (as measured by the
share of the world price going to producers) rather than to
seek full liberalization appears to have worked well in
Ghana. Given the preponderance of smallholders in the sec-
tor and the risks associated with the total withdrawal of the
government’s services, the partial liberalization experience
of Ghana’s cocoa sector has so far offered a unique example
of how it is possible to learn from past reforms and to con-
tinue to seek further reform to sustain the sector. However,
the scope for future improvements and for further learning
opportunities will require appropriate pressures from both
local political processes and from external sources.
NOTES
1. Using survey data collected in 2002, Edwin and Masters
(2003) show that the new tree varieties yield approximately
214 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
twice as much cocoa per hectare as similar-aged fields
planted with traditional trees.
2. Ghana maintains a state-owned processing plant, the
Cocoa Processing Company (CMC). Historically, CMC has
operated at low capacity. A five-year rehabilitation and
expansion program, however, allowed it to double its annual
processing capacity between 2004 and 2009.
3. The committee includes a variety of representatives
from the cocoa sector: Cocobod, government officials, and
representatives of cocoa buyers, the national cocoa farmers’
association, and haulers and transporters.
4. Indirect taxation is measured as the difference between
world prices converted using the official exchange rate and
world prices converted using the market exchange rate. Direct
taxation includes Cocobod’s marketing costs and export
duties imposed by the government (export duties have been
close to 25 percent in recent years). The share of f.o.b. prices
received by farmers does not correspond with global prices.
For example, between 1971 and 1983, the farmer share
declined sharply while global prices were rising. This period,
however, coincided with acute domestic currency overvalua-
tion in Ghana, which further eroded farmers’ real producer
prices. Similarly, in the mid-1990s, producers’ share of world
prices increased while global prices were falling.
5. How much of the program is funded by cocoa revenues
and whether any of the program is subsidized by the gov-
ernment is not clear, however.
6. In Ghana, the distinction between land ownership and
usufruct rights over what grows on land has traditionally
shaped smallholders’ investment choices.
BIBLIOGRAPHY
Abbott, P. 2002. “Towards more Socially Responsible Cocoa
Trade.” Working Paper 03-3, presented at the Annual
Meeting of the International Agricultural Trade Research
Consortium (IATRC), Monterey, CA, December 15–17.
Agrisystems Ltd. 1997. “Study of the Cocoa Sector to
Define Interventions on Behalf of Ghana’s Smallholder
with Particular Reference to the Framework of Mutual
Obligations.” Prepared for the Stablex 1992 and 1993
Allocation.
Akiyama, T., J. Baffes, D. Larson, and P. Varangis. 2001.
Commodity Market Reforms: Lessons of Two Decades.
World Bank Regional and Sectoral Studies. Washington:
World Bank.
Allman, J. M. 1993. The Quills of the Porcupine. Madison,
WI: University of Wisconsin Press.
Amanor, K. 2005. “Agricultural Markets in West Africa:
Frontiers, Agribusiness and Social Differentiation.” IDS
Bulletin 36 (2).
______. 2010. “Family Values, Land Sales and Agricultural
Commodification in Rural Ghana.” Africa 80 (1):
104–125.
Austin, D. 1978. Politics in Africa. Manchester, U.K.: Man-
chester University Press.
Awanyo, L. 1998. “Culture, Markets, and Agricultural Pro-
duction: A Comparative Study of the Investment Pat-
terns of Migrant and Citizen Cocoa Farmers in the West-
ern Region of Ghana.” Professional Geographer 50:
516–30.
Barrientos, S. W., and L. Asenso-Okyere. 2008. “Mapping
Sustainable Production in Ghanaian Cocoa: Report to
Cadbury.” Institute of Development Studies, University
of Sussex, and University of Ghana.
Bateman, M. 1965. “Aggregate and Regional Supply Func-
tions for Ghanaian Cocoa: 1946–1962.” Journal of Farm
Economics 47 (2): 384–401.
Bates, R. H. 2005. Markets and States in Tropical Africa. The
Political Basis of Agricultural Policies. Berkeley, CA: Uni-
versity of California Press.
Beckman, B. 1976. Organising the Farmers: Cocoa Politics
and National Development in Ghana. Uppsala: Scandina-
vian Institute of African Studies.
Berry, S. 1993. No Condition Is Permanent: The Social
Dynamics of Agrarian Change in Sub-Saharan Africa.
Madison, WI: University of Wisconsin Press.
______. 2009 “Building for the Future? Investment, Land
Reform and the Contingencies of Ownership in Con-
temporary Ghana.” World Development 37 (8): 1370–78.
Bloomfield, E. M., and R. A. Lass. 1992. “Impact of Struc-
tural Adjustment and Adoption of Technology on Com-
petitiveness of Major Cocoa Producing Countries.”
Working Paper 69. Paris: Organisation for Economic Co-
operation and Development.
Blowfield, M. 1993. “The Allocation of Labour to Perennial
Crops: Decision-making by African Smallholders.” NRI-
Socioeconomic Series 3, Natural Resources Institute,
University of Greenwich, Kent, U.K.
Boahene, K., T. A. B. Snijders, and H. Folmer. 1999. “An Inte-
grated Socio Economic Analysis of Innovation Adoption:
The Case of Hybrid Cocoa in Ghana.” Journal of Policy
Modelling 21 (2): 167–84.
Breisinger, C., X. Diao, S. Kolavalli, and J. Thurlow. 2008.
“The Role of Cocoa in Ghana’s Future Development.”
IFPRI Background Paper 11, IFPRI, Ghana.
Breisinger, C., X. Diao, J. Thurlow, and R. M. Al Hassan.
2009. “Potential Impacts of a Green Revolution in Africa
– The Case of Ghana.” Paper presented at the 27th IAAE
Conference, Beijing, August 16–22.
Brooks, J., A. Croppenstedt, and E. Aggrey-Fynn. 2007. “Dis-
tortions to Agricultural Incentives in Ghana.” Agricultural
CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
Distortions Working Paper 47, World Bank, Washington,
DC.
Bulíř, A. 2002. “Can Price Incentive to Smuggle Explain the
Contraction of the Cocoa Supply in Ghana?” Journal of
African Economies 11 (3): 413–39.
Coulombe, H., and Q. Wodon. 2007. “Poverty, Livelihoods,
and Access to Basic Services in Ghana.” In “Ghana Coun-
try Economic Memorandum: Meeting the Challenge
of Accelerated and Shared Growth.” World Bank,
Washington, DC.
Edwin, J., and W. A. Masters. 2003. “Genetic Improvement
and Cocoa Yields in Ghana.” Working Paper, Purdue
University, West Lafayette, IN.
Fold, N. 2000. “A Matter of Good Taste? Quality and the
Construction of Standards for Chocolate Products in the
European Union.” Cahiers d’économie et sociologie
rurales, 55–56.
______. 2001. “Restructuring of the European Chocolate
Industry and Its Impact on Cocoa Production in West
Africa.” Journal of Economic Geography 1: 405–20.
______. 2002. “Lead Firms and Competition in ‘Bi-polar’
Commodity Chains: Grinders and Processors in the
Global Cocoa-Chocolate Industry.” Journal of Agrarian
Change 2 (2): 228–47.
Fold, N., and S. Ponte. 2008. “Are (Market) Stimulants Inju-
rious to Quality? Liberalization, Quality Changes and the
Reputation of African Coffee and Cocoa Exports.” In
Globalization and Restructuring of African Commodity
Flows, ed. N. Fold and M. N. Larsen. Uppsala, Sweden:
Nordic Africa Institute.
Ghana Cocoa Marketing Board. Various years. “Cocoa Mar-
keting Board Newsletter.”
Ghana Ministry of Manpower, Youth and Employment.
2006. “Labour Practices in Cocoa Production in Ghana
(Pilot Survey).” Accra.
Gibbon, P., Y. Lin, and S. Jones. 2009. “Revenue Effects of
Participation in Smallholder Organic Cocoa Production
in Tropical Africa: A Case Study.” DIIS Working Paper
2009:06, Danish Institute for International Studies,
Copenhagen.
Gilbert, C. L. 2009. “Cocoa Market Liberalization in Retro-
spect.” Review of Business and Economics 54: 294–312.
Gill & Duffus Group. Various years. “Cocoa Statistics.”
Gockowski, J. 2007. “The Analysis of Policies, Productivity
and Agricultural Transformation in the Cocoa-Produc-
ing Rural Economies of West Africa.” STCP Technical
Report Executive Summary.
Gockowski, J., and D. Sonwa. 2007. “Biodiversity Conser-
vation and Smallholder Cocoa Production Systems in
West Africa with Particular Reference to the Western
Region of Ghana and the Bas Sassandra region of Côte
215
 d’Ivoire.” Draft paper, Institute of Tropical Agriculture,
Ibadan, Nigeria. http://www.odi.org.uk/events/2007/
11/19/434-paper-discussion-biodiversity-conservation-
smallholder-cocoa-production-systems-west-africa.pdf.
Government of Ghana. 2003. “Ghana Poverty Reduction
Strategy, 2003-2005: An Agenda for Growth and Pros-
perity.”Government of Ghana, Accra.
Gunnarsson, C. 1978. “The Gold Coast Cocoa Industry
1900–1939. Production, Prices and Structural Change.”
PhD thesis, Department of Economic History, Lund Uni-
versity, Lund, Sweden.
Gyimah-Brempong, K. 1987. “Scale Elasticities in Ghanaian
Cocoa Production.” Applied Economics 19: 1383–90.
Hattink, W., N. Heerink, and G. Thijssen. 1998. “Supply
Response of Cocoa in Ghana: A Farm-Level Profit
Function Analysis.” Journal of African Economies 7 (3):
424–44.
Herbst, J. 1993. The Politics of Reform in Ghana, 1982–1991.
Berkeley, CA: University of California Press.
Hill, P. 1963. The Migrant Cocoa Farmers of Southern Ghana.
A Study in Rural Capitalism. Cambridge, UK: University
Press.
International Cocoa Organization. Various years. “Quarterly
Bulletin.” London.
Konings, P. J. J. 1986. The State and Rural Class Formation in
Ghana: A Comparative Analysis. Leiden, Netherlands:
African Studies Centre.
Laven, A. 2007. “Marketing Reforms in Ghana’s Cocoa Sector:
Partial Reforms, Partial Benefits?” Background Note,
Overseas Development Institute, London. http://www.odi.
org.uk/resources/download/420.pdf.
Masdar Ltd. 1998. Socio-Economic Study of the Cocoa Farm-
ing Community. Wokingham, United Kingdom.
McKay, A., and H. Coulombe, 2003. “Selective Poverty
Reduction in a Slow Growth Environment: Ghana in the
1990s.” Human Development Network, World Bank,
Washington, DC.
McMillan, M. 1998. A Dynamic Theory of Primary Export
Taxation. Discussion Paper 98-12. Department of Eco-
nomics, Tufts University.
Norris, K. 2008. “Carbon, Biodiversity and Cocoa Farming
in Ghana.” Climate Change Unit, Forestry Commission
of Ghana, Accra.
Nyanteng, V. K. 1993. “The Prospect of the Ghanaian Cocoa
Industry in the 21st Century.” Paper presented at the
International Conference on Cocoa Economy, Bali,
Indonesia, October 19–22.
Obiri, B. D., G. A. Bright, M. A. McDonald, L. C. N. Anglaaere,
and J. Cobbina. 2007. “Financial Analysis of Shaded Cocoa
in Ghana.” Agroforestry Systems 71: 139–49.
216 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
Okali, C. 2010 “The Organisation of Cocoa Production on
the Farm and its Evolution over Time.” Unpublished
paper.
Opoku, E., R. Dzene, S. Caria, F. Tea, and A. Zeitlin. 2009.
“Improving Productivity through Group Lending:
Report on the Impact Evaluation of the Cocoa Abrabopa
Initiative.” Centre for the Study of African Economies,
University of Oxford, Oxford, United Kingdom. http://
www.csae.ox.ac.uk/output/reports/pdfs/rep2008-01.pdf.
Piasentin, F. and L. Klare-Repnik. 2004. “Quality Matters.”
Global Research on Cocoa and CABI, Conservation
International.
Quisumbing, A. R., E. Payongayong, J. B. Aidoo, and K.
Otsuka. 2001. “Women’s Land Rights in the Transition to
Individualized Ownership: Implications for the Manage-
ment of Tree Resources in Western Ghana.” Economic
Development and Cultural Change 50 (1): 157–81.
Rourke, B. E. 1974. “Profitability of Cocoa and Alternative
Crops in Eastern Region, Ghana.” In Economics of Cocoa
Production and Marketing, ed. A. Kotey, C. Okali, and
B. E. Rourke. Proceedings of Cocoa Economics Research
Conference, University of Ghana, Legon, 1973.
______. 2007a. “The Cocoa Sector: Expansion, or Green
and Double Green Revolutions?” Overseas Development
Institute Background Note, London.
______. 2007b. “International Perspectives on the Cocoa
Sector: Expansion or Green and Double Green Revolu-
tions?” http://www.odi.org.uk/events/2007/11/19/434-
presentation-session-1-international-perspectives-cocoa-
sector-francois-ruf.pdf.
______. 2009. “Libéralisation, cycles politiques et cycles du
cacao: le décalage historique Côte-d’Ivoire-Ghana.”
Cahiers Agricultures 18 (4).
Ruf, F., and K. Burger. 2001. “Planting and Replanting Tree
Crops. Smallholders’ Investment Decision.” Unpublished
paper, Centre de coopération internationale en recherche
agronomique pour le développement (CIRAD), Paris.
Ruf, F., and H. Zadi. 1998. “Cocoa: From Deforestation to
Reforestation.” Prepared for the First International
Workshop on Sustainable Cocoa Growing, Smithsonian
Institute, Panama.
______ and P. S. Siswoputranto. 1995. Cocoa Cycles: The
Economics of Cocoa Supply. Cambridge, U.K.: Woodhead
Publishing Ltd.
Shepherd, A. W., and S. Farolfi. 1999. “Export Crop Liberal-
ization in Africa: A Review.” Agricultural Services Bul-
letin 135, FAO, Rome.
Sjaastad, E. and D. Bromley. 1997. “Indigenous Land Rights
in Sub-Saharan Africa: Appropriation, Security and
Investment Demand.’” World Development 25: 549–62.
Steedman, C. 2003. “Agriculture in Ghana: Some Issues.”
Unpublished paper, World Bank, Washington, DC.
Stryker, J. D. 1990. “Trade, Exchange Rate, and Agricultural
Policies in Ghana.” World Bank Comparative Studies,
World Bank, Washington, DC.
Takane, T. 2002. The Cocoa Farmers of Southern Ghana:
Incentives, Institutions, and Change in Rural West Africa.
Chiba, Japan: Institute of Developing Economies and
Japan External Trade Organization.
Teal, F., A. Zeitlin, and H. Maamah. 2006. “Ghana Cocoa
Farmers Survey 2004: Report to Ghana Cocoa Board.”
Centre for the Study of African Economies, University of
Oxford, Oxford, United Kingdom.
Vigneri M. 2005. “Trade Liberalisation and Agricultural Per-
formance: Micro and Macro Evidence on Cash Crop Pro-
duction in Sub-Saharan Africa.” DPhil thesis, University
of Oxford, Oxford, United Kingdom.
______. 2008. “Drivers of Change in Ghana’s Cocoa Sector.”
IFPRI-GSSP Background Paper 13, International Food
Policy Research Institute, Washington, DC.
CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY
Vigneri M., and P. Santos. 2008. “What Does Liberalization
without Price Competition Achieve? The Case of Cocoa
Marketing in Rural Ghana.” IFPRI-GSSP Background
Paper 14. International Food Policy Research Institute,
Washington, DC.
Vigneri M., F. Teal, and H. Maamah. 2004. “Coping with
Market Reforms: Winners and Losers among Ghanaian
Cocoa Farmers.” Report to the Ghana Cocoa Board,
Accra.
World Bank. 2007a. World Development Report: Agriculture
for Development. Washington, DC: World Bank.
______. 2007b. “Ghana: Meeting the Challenge of Acceler-
ated and Shared Growth.” Country Economic Memoran-
dum, World Bank, Washington, DC.
______. 2008. “Country Brief: Ghana.” World Bank, Wash-
ington DC.
Zeitlin, A. 2006. “Market Structure and Productivity Growth
in Ghanaian Cocoa Production.” Unpublished paper,
Centre for the Study of African Economies, University of
Oxford, Oxford, U.K.
217
 CHAPTER 13
Apparel Exports in Lesotho: The State’s
Role in Building Critical Mass for
Competitiveness
Mallika Shakya
esotho, a country of 1.9 million people, has been
L
quite successful in capitalizing on the African
Growth and Opportunity Act (AGOA), the eco-
nomic partnership agreement between the United States
and Africa launched in 2000 to increase African exports to
the U.S. market. Lesotho’s experience with its apparel
exports demonstrates that competitiveness is not only
about firm-level costing but about developing a broader
coordination mechanism through which the government
delivers public goods necessary for overcoming market fail-
ures in such a way that private sector investment is trig-
gered and sustained.
A notable aspect of Lesotho’s experience is the govern-
ment’s decision to combine an apparel industry competi-
tiveness initiative with a series of early-stage incentives for
investors—a tactic that has yielded positive results for
Lesotho thus far. The sequencing and balancing of such pol-
icy measures has differentiated Lesotho’s apparel export
path from those of other African countries in recent years.
Its proximity to South Africa and its membership in the
Southern African Customs Union offers Lesotho unique
geographic and fiscal advantages that other countries may
not have, but the industrial policy measures and institu-
tional reforms undertaken by the government of Lesotho
are relevant for several African countries. Lesotho’s success
can especially be attributed to its direct engagement with
global and regional apparel buyers and an overall policy
acknowledgement of the rapidly changing structure of the
Mallika Shakya is Wolfson Research Fellow at the School of Interdisciplinary Area Studies, University of Oxford.
global apparel trade. Over time, Lesotho has started to climb
the apparel value chain and overcome issues of corporate
social responsibility.
To better understand the policy measures behind the
success factors involving Lesotho’s apparel trade, this chap-
ter asks three main questions. First, how are spatial and
economic factors balanced in pursuing a coherent compet-
itiveness strategy in Lesotho’s apparel sector? Second, what
were some of the industry-specific strategies employed by
the government of Lesotho in overcoming the government
and market failures? How are they complemented with
economywide measures and socioeconomic measures? And
third, what public-private collaborative efforts have led to
reforms that maximized private sector investments? How
have bottlenecks related to transportation, logistics, innova-
tion, and quality been overcome to produce the success of
the apparel industry in Lesotho?
In exploring these questions, this chapter takes as a basic
premise that international trade is no longer a random phe-
nomenon where anonymous producers and buyers engage
in exchange of what they produce and consume. Rather,
international trade and investment decisions increasingly
depend on a successful ecosystem of producers, suppliers,
and providers of specialized services and infrastructure,
along with cooperation from the public sector on essential
issues of tax, finance, innovation, and quality control.
There is a growing tendency for global apparel investors,
buyers, and producers to rely on concerted public support
219
 that harnesses new production methods, facilities, and
logistical models. Beyond macroeconomic soundness,
countries should also give attention to micro- and meso-
level issues related to delivery of necessary public goods for
private investment. All of these factors add up to a better
business-enabling environment necessary for sustained
competitiveness.
Not only is the apparel industry an entry point for a
broad range of light manufacturing industries but, if man-
aged well, it has the capacity to unleash the growth and
competitiveness potential of an economy. Growth of the
apparel industry makes immediate contributions to
employment and creates tremendous potential for back-
ward and forward linkages that are important to developing
countries such as Lesotho.
Lesotho exported just under $350 million worth
of apparel to the United States in 2008, approximately
29 percent of apparel exports (Edwards and Lawrence
2009) from all of Sub-Saharan Africa to the United States
for that year (figure 13.1). Although this marks a dip from
the amount exported in 2004, Lesotho still exports sub-
stantially more apparel than its closest rivals in the sector,
Madagascar and Kenya. The value of Lesotho’s apparel
exports to the United States is about three times higher
than those of Swaziland, a country that has similar geopo-
litical conditions and that follows Lesotho’s growth path to
a certain extent. Lesotho is also far ahead of neighboring
Mozambique, which exported less than $400,000 worth of
apparel to the United States in 2008, approximately 0.1 per-
cent of the amount Lesotho exported.
Other African countries, such as Cameroon, Ghana,
and Tanzania, lag far behind Lesotho in apparel exports,
Figure 13.1 U.S. Imports of Clothing and Textiles from
AGOA Countries
500
450
400
350
$ millions
and even middle-income countries in Sub-Saharan Africa
such as Mauritius and South Africa are rapidly losing their
comparative and competitive advantages in the apparel
sector. Although both these countries exported amounts
comparable with that of Lesotho in 2001, the situation
changed dramatically in 2008, when Mauritius exported
only about $100 million worth of apparel and South Africa
exported a negligible amount to the United States. While
South Africa exports some of its apparel to the European
Union (EU) and has a significant domestic market, it is
quite striking that Lesotho, a country with a very small
domestic market, exports three times as much apparel than
Mauritius and is decisively ahead of larger African countries
Madagascar and Kenya. Even though the global financial
crisis negatively affected Lesotho’s apparel exports in 2008
and 2009, it is difficult to challenge the country’s success
between 2001 and 2007.
Figure 13.1 shows that Lesotho’s apparel exports to the
American market grew exponentially in the late 1990s and
early 2000s, before more or less stabilizing in recent years. A
brief slowdown due to the uncertainty in international trade
agreements resulted in the closure of six firms and a mod-
erate compromise on exports in 2004 and 2005, but four
new firms entered the market at the same time, and there
was a significant increase in total employment in the apparel
industry. Despite the impact of the 2008/2009 global finan-
cial crisis recovery has been relatively quick.
The magnitude of Lesotho’s apparel industry success
becomes clearer when its export figures are weighed against
the size of its population. While high population density
and availability of cheap labor explain the success of the
apparel industry in Bangladesh, China, and India, the situa-
tion is not that straightforward in Africa. As measured by
apparel exports per capita, smaller countries with lower
population density are outperforming larger countries with
large labor forces (figure 13.2). Lesotho, in fact, outranks
other apparel-producing countries in Sub-Saharan Africa,
with apparel exports per capita of $177 as of 2009. Swazi-
land and Mauritius have the second- and third-largest per
capita apparel exports, respectively, among Sub-Saharan

300 African countries.

250
200 The apparel industry represents a significant proportion
150 of Lesotho’s gross domestic product (GDP), trade, and
100
50 employment. Clothing, in fact, has represented the majority
0
of Lesotho’s total exports since the late 1990s (figure 13.3).
19 9
90
91

19 2
19 3
19 4
95

19 6
97

19 8
99

20 0
20 1
02

20 3
20 4
05

20 6
07
08
8

9
9
9

0
0

0
0

0
19

19
19

19

19

20

20

20

20

In addition to Lesotho’s successful entry into the U.S.

Lesotho Madagascar Kenya
apparel market, investors are producing higher-value-
Mauritius Swailand South Africa
added items such as woven shirts and men’s trousers and
Source: U.S. Trade Development Agency. are now targeting the regional (South African) and EU

220 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS
Figure 13.2 Per Capita Apparel Exports to the Figure 13.4 Employment Trends in Lesotho
United States
60,000
200 50,000

Number of workers
177
40,000
150 30,000
20,000
95
100
$

80 10,000
0
50

00

01

02

03

04

05

06

07
20

20

20

20

20

20

20

20
15
0.1 0.2 0.7 5 7 All other Textiles and clothing Total
0
ia

nd

o
w

iu
ca
ric

an

ny

Source: Lesotho Bureau of Statistics.

th
op

ala

ila
rit
as
w

Ke

so
Af
hi

az
M

au
ts

ag

Le
Et

Sw
Bo

ad

M
ut

M
So

Source: Apparel data, Edwards and Lawrence 2009; population data, United GLOBAL AND REGIONAL TRADE AGREEMENTS
Nations database. SET THE STAGE FOR APPAREL TRADE

AGOA, which went into effect in October 2000, designated

34 “least developed” countries in Sub-Saharan Africa as eli-
Figure 13.3 Composition of Lesotho’s Exports gible for specialized trade benefits in the United States.1
Swaziland was designated as the 35th AGOA-eligible coun-
100
try in January 2001. Under AGOA, Lesotho’s apparel exports
80 enjoy duty-free access to the American markets until 2012,
making them up to 15 percent cheaper in the U.S. market
60
Percent

than exports from non-Sub-Saharan African countries. And

40 because Lesotho is still categorized as a least developed
country under AGOA, it is exempt from the rules of origin
20 clause, meaning that it is free to import fabrics and other
0 inputs from anywhere in the world.
Because it was among the first countries to put necessary
98

99

00

01

02

03

04

05

06

07
19

19

20

20

20

20

20

20

20

20

systems into place, Lesotho was among the first African

Clothing Diamonds
Footwear All other goods
Source: Lesotho Bureau of Statistics.
markets. Lesotho has also emerged as one of the few African
countries where regionally grown cotton is processed into
denim and exported as part of final apparel products to the
United States.
Beyond economic considerations, direct employment
benefits make Lesotho’s apparel industry an important part
of the national economy. In 2010 just under 40,000 of a total
population of 1.9 million people were employed in the
apparel industry. Figure 13.4 shows the extent to which
labor conditions in the apparel sector dictate the overall
employment conditions in Lesotho.
Television sets
countries to reap the benefits of AGOA. It was not until
February 2002, for example, that neighboring Mozambique
was technically qualified for the U.S. apparel provision. To
qualify for the AGOA, each country must ensure the United
States Trade Representative’s office that it has the adminis-
trative capacity to effectively monitor against possible abuse
of the privileges granted under AGOA.
The period from 2000 to 2004 was a unique time for
African countries seeking to seize opportunities in the
global apparel market. During these years African countries
not only received preferential treatment in the U.S. market
under AGOA, but the preference came at a time when
quota-based restrictions were being imposed on exports
from large Asian countries to the United States under the
Multifibre Agreement (MFA).2 For some large Asian coun-
tries, the MFA-imposed quota meant that they could only
export less than half of what they had historically exported.

CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS 221
 The MFA quota, however, could be borrowed from the
forthcoming year and any unused quota from one year
could be carried over to the next. More importantly, quotas
could be transferred between countries (that is, countries
wishing to export more could purchase quotas from coun-
tries not using them). At times the value of quota reached
up to 30 percent of the original value of apparel being
traded. Over time, this quota system encouraged large
Asian manufacturers to relocate their operations to less-
developed countries, such as Lesotho, that could participate
in preferential trade arrangements with the United States. In
addition to the tariff-based preferences it offered, Lesotho’s
historical ties to East Asian countries made it attractive as a
location for apparel manufacturing.
KEY FACTORS BEHIND LESOTHO’S
COMPETITIVENESS IN APPAREL
There is no set formula to explain why some countries are
more competitive than others in a given industry. In general,
competitiveness stems from sophisticated firms engaging
in productive ventures, supported by a policy and institu-
tional environment in which firms are able and encouraged
to forgo rent-seeking behavior and government and market
failures are overcome. In this regard, competitiveness is a
kind of continued filtration process through which well-
performing firms and markets distinguish themselves from
weaker firms.
Although a sound macroeconomic environment is nec-
essary for increased competitiveness of a sector, that alone
will not create competitiveness when the basic building
blocks of industrialization, such as having a critical mass of
firms that can develop a sustainable economic ecosystem,
are missing, as is often the case in Sub-Saharan Africa.
Development of a critical mass of firms, however, is subject
to public support for necessary institutional and infra-
structural mechanisms. In the case of Lesotho’s apparel
industry, the institutional framework was created through
a coordinated public-private partnership.
To answer the question of why Lesotho succeeded in
apparel while most other African countries failed, despite
having equally preferential market access through AGOA, a
closer examination of the costs of apparel manufacturing is
needed. Beyond having the appropriate macroeconomic fun-
damentals in place, success in apparel manufacturing is
related to labor and overhead costs, which together represent
the bulk of the cost of apparel manufacturing. Figure 13.5
shows that while labor makes up as much as 49 percent of
the nonmaterial costs of apparel manufacturing, overhead
222 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS
Figure 13.5 Breakdown of Nonmaterial Costs of Apparel
Production in Lesotho
Overhead/
Labor Profit Misc
utilities
(49%) (10%) (20%)
(20%)
0 20 40 60 80 100
Percent
Source: World Bank 2006.
and utilities are 20 percent of such costs. Both of these costs
are dependent on the presence of public goods that are nec-
essary for factories’ efforts to reduce cost and enhance com-
petitiveness.
Factory market
A typical apparel factory in Southern Africa spends approx-
imately 20 percent of total production costs in overhead
and utilities, including land, electricity, and transportation.
Land ownership in Southern Africa was historically under
a communal system, and later a public system, hence there
is no commercial land market. Availability of land is a
major problem for any foreign investor seeking to start a
business venture in Southern Africa. In Lesotho, even if
land becomes available, the cost of renting it is sometimes
as high as in developed countries, and obtaining a permit
to construct a factory or other commercial building on the
land takes an average of 16 months. According to the Doing
Business report, it takes 601 days to obtain a construction
permit in Lesotho, compared with the Sub-Saharan African
average of 260; and 157 days in developed countries. The
cost of obtaining a construction permit in Lesotho is 670
percent of per capita gross national product, compared
with 56 percent in developed countries, while the unsubsi-
dized rate of renting a factory shell in Lesotho is more than
twice the rate for a comparable facility in South Africa and
other developed countries.
While national land ownership systems are not likely to
be reformed overnight, it seems logical for governments to
offer transitional arrangements for commercial land use in
the form of hassle-free leasing contracts. This point was
highlighted in a study commissioned by the government of
Lesotho in 2002 to develop a comprehensive apparel indus-
try development plan. Among the 11 principal constraints
to the growth of the apparel industry in Lesotho identified
in the report, lack of industrial land and factory shells were
the top two constraints.
To compensate for high land costs, the government of
Lesotho constructed industrial zones and serviced factory
shells available for rent by foreign investors, allowing
investors to avoid cumbersome bureaucratic processes. Two
large industrial zones—Maputsoe on 29 hectares of land
and Ha on 10 hectares—were built between 1991 and 1995.
Additional zones were built subsequently on the
Lesotho–South Africa border areas of Mohale’s Hoek,
Mfeteng, and Butha Buthe. Currently these fully functional
industrial zones in Lesotho together house about 60 facto-
ries and employ just under 40,000 workers (figure 13.6).
The industrial zones provide a range of real estate ser-
vices to assist manufacturing and commercial enterprises,
including fully serviced industrial plots, customized factory
buildings, and development of commercial properties for
leasing. The buildings are fully serviced with electricity, tele-
phone, water, and sewerage. All six zones can be reached by
tarred road, are within reach of the public transport system,
and have access to rail links connecting them to apparel dis-
tribution hubs in Johannesburg.
Because the land tenure system in Lesotho does not pro-
vide for land being bought or sold, investors were offered
long-term leases within these zones and were given the
option to sublease if necessary. Rental procedures and reg-
ulatory processes were simplified and centralized so that
the export promotion agency, Lesotho National Develop-
ment Corporation (LNDC), could undertake all necessary
administrative procedures and cater to the business needs
of potential foreign investors. In fact, LNDC’s success as
an export promotion agency draws enormously on its suc-
cess as a property developer and service provider, which
catalyzed development of a critical mass of first-generation
apparel manufacturing companies in Lesotho. After man-
aging these services in house for several years, LNDC
recently began subcontracting day-to-day property opera-
tion and maintenance tasks to JHI Real Estate, a prominent
estate management company active in several other African
countries.
Timing was crucial to the success of industrial zone
construction in Lesotho, a point that is often omitted from
policy discussions. After AGOA became effective in 2000,
and foreign investors operating in East and South Asia

Figure 13.6 Industrial Zones in Lesotho

IBRD 38607
MAY 2011
28° 29°
SOUTH AFRICA Butha-Buthe INDUSTRIAL PRODUCTION ZONES
BUTHA- DISTRICT CAPITALS
Leribe BUTHE
29° LERIBE 29° NATIONAL CAPITAL
Teyateyaneng
BEREA MOKHOTLONG DISTRICT BOUNDARIES
Mokhotlong
MASERU INTERNATIONAL BOUNDARIES

MASERU Thaba-Tseka
THABA-TSEKA
MAFETENG
Mafeteng

30° MOHALE’S HOEK QACHA’S NEK 30° LESOTHO

Qacha’s Nek
Mohale’s Hoek
SOUTH AFRICA
QUTHING
Quthing 0 50 Km.
0 30 Mi.
29° 29°

Source: Government of Lesotho.

CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS 223
 began to operate factories in Africa, and Lesotho was
among the few countries that already had industrial zones
with well-equipped and serviced factory shells. As a specific
response to AGOA, Lesotho offered the apparel investors
subsidized rent for at least the first five years of operation.
While the sustainability and macroeconomic rationale of
this policy decision may be questioned, it virtually pre-
empted the efforts of other countries in the region to
attract foreign direct investment (FDI).
Export and investment promotion
Observers often claim that export and investment promo-
tion activities have not yielded results and that they have
instead become rent-seeking activities. This is true in sev-
eral cases both in Africa and elsewhere in which export
promotion agencies are captive to poor leadership due to
heavy government involvement, inadequate funding,
cumbersome bureaucracy, and lack of client orientation.
At times, countries suffer from strong antitrade bias and
anti-FDI sentiment. By many measures, Lesotho’s success
in export and investment promotion defies the general
trend in developing countries. The LNDC is the main
wing of the government charged with the implementation
of export and investment promotion programs. Founded
in 1967 by an act of parliament, LNDC was amended in
1990 and again in October 2000. LNDC is primarily
owned by the government of Lesotho (90 percent); DEG,
a German finance company, owns the remaining 10 per-
cent. The LNDC is part of the Ministry of Trade and
Industry, Cooperatives and Marketing, which is responsi-
ble for overall policy direction on industrialization in
Lesotho.
The LNDC works with medium and large companies in
the manufacturing and services sectors to strengthen
Lesotho’s exports and generate foreign investment in
Lesotho. Rather than relying on a broad strategy to pro-
mote all exports, the LNDC focuses on nontraditional
exports. Because of Lesotho’s extreme concentration in
apparel, the LNDC also has functioned as a de facto gar-
ment promotion agency in the past several decades, devel-
oping with it several sector-specific skills and insights. Such
insights have proved instrumental in ensuring that incen-
tive packages offered to investors are not poorly utilized or
captured by rent-seekers.
In addition to the LNDC’s role as real estate developer,
facilitator, and service provider, it acts as a one-stop shop
that covers most aspects of company operations, ranging
from business inquiries, site selection, firm registration,
224 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS
regulation and facilitation on various aspects of production
and sales, and management of government incentive
schemes and tax procedures (box 13.1). Through its Invest-
ment Promotion Center (IPC), the LNDC markets Lesotho
as a destination of choice to potential manufacturers. To
achieve this, it has developed an information dissemination
system that reaches out to potential investors within
Lesotho, Southern Africa, and beyond. The LNDC has sent
trade missions to South and East Asia to mobilize networks
of firms that are already invested in Lesotho. One factor
behind the success of Lesotho’s export and investment pro-
motion is that the IPC has adopted tailor-made strategies to
attract and sustain investments from very different
investors, including those from South Africa; China; Tai-
wan, China; and elsewhere. In recent years, the IPC has also
sought to attract yarn, trim, zipper, button, bag, hanger, and
carton manufacturers to Lesotho. To some extent, these spe-
cialized manufacturers can assist Lesotho in taking full
advantage of AGOA and allow a structured approach to
inward investment.
Labor coordination and skill development
At 49 percent, labor is the largest nonmaterial cost in
Lesotho. It goes without saying that labor costs directly
influence investors’ decisions about where to locate their
operations and buyers’ decisions about where to choose
suppliers. In the case of Lesotho, however, labor costs are
not the primary point of consideration for investors and
buyers. In other words, investors have chosen Lesotho
despite its relatively higher labor costs compared with
other Sub-Saharan African countries. Nonetheless, because
production and sourcing dynamics are constantly chang-
ing in the global apparel market, it is important that
Lesotho takes measures to reduce labor costs while
improving labor productivity.
Labor costs are high in Lesotho compared to other
Sub-Saharan African countries. Hourly wages range
from approximately $0.14 to $1.80 in apparel-producing
African countries as well as in several non-African coun-
tries with significant apparel sectors such as Bangladesh,
China, and Sri Lanka. Wages in Lesotho average $0.46 an
hour, lower than those in Mauritius and South Africa but
significantly higher than those in Ethiopia, Kenya, Mada-
gascar, Mozambique, and Swaziland. Even when productiv-
ity is taken into account (labor costs per shirt), Lesotho still
is an expensive manufacturer compared with Ethiopia,
Kenya, Madagascar, and Mozambique (figure 13.7).
 Box 13.1 LNDC as a One-Stop Shop
Company support
■ Business registration procedures
■ Acquisition of permits and manufacturing licenses
■ License and residency paperwork for foreign work-
ers, managers, and owners
■ Arrangement of site visits and assistance in selection
of suitable sites
■ Key focal point for contact with relevant ministries
on business regulations
■ Facilitation of contact with business companies sup-
plying services
■ Facilitation of skill and technology development
programs
■ Industrial relations if disputes arise with workers
Management of incentive schemes
■ Unimpeded access to foreign exchange
■ Export finance facility
Source: Consultations with LNDC officials.
Labor unions in Lesotho have undergone a major
transition. The first apparel factories were set up in
Lesotho by South African companies in the 1980s, and were,
for the most part, subsidiaries of bigger industrial conglom-
erates that had specialized for decades in the South African
market. Primarily located in Maputsoe, these companies
initially established higher labor standards for Lesotho.
They negotiated recognition agreements with the trade
union of employees at the LNDC and the Department of
Labor. While those negotiations were frequently difficult
and protracted, they inevitably resulted in pay increases for
workers. Over the years, negotiations with trade unions
became commonplace among apparel companies in
Lesotho, even among those where the majority of employ-
ees were not union members.
The new generation of apparel manufacturers and buy-
ers that moved to Lesotho in response to the MFA and
AGOA provisions created a demand for a very different set
of labor skills and expectations. In short, the industry
became less regulated by local labor codes and more influ-
enced by the global momentum around labor-related cor-
porate social responsibility. The apparel industry in Lesotho
underwent a major labor crisis in the 2000s as a result, one
that involved policy deadlocks and street riots. On one
CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS
■ Short- and long-term loans
■ Import value added tax (VAT) credit facility for local
purchase of raw materials and capital goods
Tax management oversight
■ No tax on income generated from exports outside
the Southern African Customs Union
■ Permanent maximum manufacturing tax rate of 10
percent on profits
■ No tax on dividends to local or foreign shareholders
■ Free repatriation of profits
■ Double taxation agreements with Germany, Mauri-
tius, South Africa, and the United Kingdom
Investment attraction
■ Information dissemination to targeted investor pools
■ Trade missions to targeted markets and investor
pools
hand, workers were upset that job stability was no longer
guaranteed; on the other, factory owners were frustrated
about low labor productivity. Fueling the antagonism were
several cases of exploitation of workers by foreign factory
owners and foreign supervisors.
The Lesotho Clothing and Allied Workers Union
(LECAWU) is the predominant union representing employ-
ees of the apparel industry. It was formed under the terms
of the national labor code of Lesotho in1992 and is recog-
nized by the government of Lesotho as the representative of
labor. Historically, anti-union sentiment has made it diffi-
cult for LECAWU to organize workers; the unwillingness of
management, particularly in Maseru and Thetsane, to
engage in meaningful negotiation also impeded the union.
Nevertheless, LECAWU membership has grown in recent
years as a result of AGOA provisions, changes in apparel
companies’ codes of conduct, and increasing interest in
labor relations among company owners and managers. The
labor movement in Lesotho encouraged the government to
create the legislative framework that set up the Industrial
Relations Council, the Directorate for Dispute Prevention
and Resolution, and the Labor Appeal Court. This impor-
tant piece of legislation provides for the settling of labor dis-
putes through mediation, conciliation, and arbitration,
225
 Figure 13.7 a. Factory-Floor Productivity and Labor Costs in Apparel Assembly; b. Hourly Wages for Apparel Assembly
(US cents)

a b
Labor Ethiopia 14
Daily cost per Ghana 18
Country productivity shirt ($)
Bangladesh 21
Mozambiquea 25
Mozambiquea 10–11 .16 Pakistan 26
Ghana 12 .12 Madagascar 26
Ethiopia 10–12 .12 India 27
Kenya 12–15 .18 Kenya 30
Madagascar 14–15 .16 Nepal 31
Lesotho 18 .19
Swaziland 33
South Africa 15 .65
Sri Lanka 35
India 16 .17
Lesotho 46
Chinab 18–22 .29
EPZ China 72
Mauritius 125
South Africa 179.4

Source: Labor productivity figures and costs are from Eifert, Gelb, and Ramachandran (2005) except for those for Ethiopia, which were calculated using
Global Development Solutions (2005). Hourly wages were collected from government sources.
Note: Daily productivity is the number of men’s casual shirts made by a machine operator in one day.
a. Mozambique wage indicates the minimum industry wage specified under the national labor law.
b. Data for China is for export processing zones only.

effectively creating an independent authority that is repre-
sentative of the government, employees, and employers.
The global labor movement has benefited the local labor
movement in Lesotho. As manufacturing has shifted from
developed to developing countries, there has been growing
concern among developed-country consumers, as well as
from activists and nongovernmental organizations (NGOs),
regarding the wages and work conditions under which prod-
ucts are manufactured. As a result, most major multinational
companies have adopted codes of conduct they require their
suppliers to meet. In addition, a number of NGOs have
devised codes of conduct in partnership with the multina-
tionals: these include the U.K.-based Ethical Trading Initia-
tive (sponsored, in part, by the Department for International
Development, or DFID), the Institute of Social and Ethical
Accountability, the Global Reporting Institute, the Nether-
lands-based Clean Clothes Campaign, the U.S.-based Fair
Labor Association, the U.S.-based Worker Rights Consor-
tium, and the EU’s European Code of Conduct. A major
private initiative called Social Accountability 8000
(SA8000), along with the nonprofit organization Account-
Ability’s AA1000, are attempting to create a benchmark for
social accountability much like the ISO9000 has done for
quality. Meanwhile, several multinational apparel manufac-
turing companies have formed the Worldwide Responsible
Accredited Production (WRAP) principles to accredit sup-
pliers. In addition, all major multinational apparel manu-
facturing companies have their own codes of conduct.
The need to comply with the various codes of conduct
has created a major change in most apparel factories in
Lesotho. And though the inspections surrounding such
codes of conduct can measure a degree of compliance,
industrial relations will be smooth and sustainable in the
long term only if they contribute to productivity.
Labor productivity must increase. A typical apparel
factory now works with its workers, regulators, and service
providers as a group to drive down costs and improve
standards. By working with providers of training and
skills, factories are able to shift to new production meth-
ods that reduce costs, improve product quality, and
enhance factories’ commitment to their workers. Nike’s
shift from a normal assembly line production system to
“lean manufacturing,” for example, has been marked with

226 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS
significantly higher productivity and better labor condi- of machinists, its primary aim of launching entrepreneurs
tions. The initiative has been followed by the incorporation into the formal garment subsector was not successful, and
of more sophisticated production planning and merchan- it was terminated in 2003. Lesotho’s experience with the
dising systems, such as “Just in Time,” that aim to eliminate LGC is similar to that of several other developing coun-
inefficiencies in space, time, and activity and also manage tries, where the public sector has developed channels to
the inventory stock. support the private sector that turn out to be ineffective.
Most buyers of apparel assess their producers on four key Coordination and incentive structures have been major
criteria: cost, quality, timeliness, and corporate responsibility problems in making public sector efforts useful for the pri-
(that is, labor and environmental standards) (figure 13.8). vate sector.
After an initial baseline is developed, buyers regularly mea- No other training institutes have been created to fill the
sure factories’ performance to track delivery, quality, and void left by the termination of the LGC. Several other train-
prices over time. Indicators are typically expected to improve ing establishments, such as Lerotholi Polytechnic and the
each quarter. Such proactive measures toward labor produc- Institute of Development Management, offer a range of
tivity are not yet under way in Lesotho, although a number generic management and general studies, although none are
of initial steps have been taken. The fact that Lesotho is at the tailored to the needs of the formal garment sector. Lerotholi
lower end of productivity in apparel manufacturing means Polytechnic offers a two-year series of courses related to the
that it needs to develop systematic analyses to diagnose bot- apparel industry, for example, but the focus is on design,
tlenecks and then design institution measures to tackle those pattern construction, and bespoke tailoring. The Commer-
bottlenecks accordingly. cial Training Institute offers similar courses that lead to the
award of a certificate in tailoring and dressmaking. The
Successful skill development initiatives. In 1999 Basotho Enterprise Development Corporation (BEDCO)
the Lesotho Garment Center (LGC) was opened to provide also runs a series of short courses, including marketing skills
necessary training for the apparel workforce. Developed and, actively seeks to identify and focus on the future needs
with funding from the DFID, the LGC had the objective of the garment industry. Saint Luke’s Mission in Maputsoe
of training workers and entrepreneurs in the manage- offers a three-year course providing students with a range of
ment of industrial apparel production in anticipation of apparel industry skills, including pattern construction,
their participation in the opportunities presented by machine knitting, small business studies, leadership, man-
AGOA. Although the LGC succeeded in training hundreds agement skills, and industrial garment production on the
appropriate machinery.
Upon completion of these courses, many students are
Figure 13.8 Buyer’s Scorecard for Assessment of employed in the local craft industry or their own enter-
Producer Competitiveness prises, but very few graduates from these institutes end up
working on the apparel shop floors. The training institutes
are seldom approached by apparel manufacturers to train
Gross margin Quality
n Product defective rate
managers, workers, or technicians, underscoring the need
n Labor productivity against target 0.9% for sector-specific contents and modules in training insti-
n Scale n Vendor test submit rate tutes for them to be truly useful to employers. This scenario
against target 10.0% for use
n Locational advantage
of restricted substance in highlights a larger problem in the apparel industry in
n Connectivity
production, finishing and Lesotho: while existing training institutes do not cater to the
packaging
apparel industry, apparel factories report that their in-house
Corporate
training facilities are not sufficient. For the most part,
Delivery responsibility trainees in factories are taught single machine skills rather
GAC: Goods at consolidator, ‘C’ rated factories: lack of
than being cross-trained.
the date and the product basic terms of employment
leaves the factory and legal non-compliance Most apparel factories in Lesotho also acknowledge the
OTP: On time receipt to the ‘D’ rated factories: unwillingness need for trained supervisors and indicate that they are
final port after dispatching, to comply with CR standards
and denial of access to
prepared to pay for it. In the factory survey, 69 percent of
loading and shipping
inspectors interviewees confirmed that they would support training
programs that addressed their particular sector needs if such
Source: World Bank 2006. programs were delivered in a way that minimized employees’

CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS 227
 time away from the workplace. Approximately 11 percent
said they might support such a facility, while 20 percent said
they would not. Only one of the companies surveyed had
actually developed a supervisory training course, and is
using its own staff to deliver the training. The course covers
a range of topics, from organization and planning to indus-
trial relations and health and safety. The most common rea-
sons companies gave for developing their own courses were
that no such courses were available locally and that they had
concerns about quality of delivery and training materials
provided by third parties.
The fact that services offered by existing training insti-
tutes are not used by apparel factories while the same
factories acknowledge need for more effective training
indicates a coordination gap. While an initiative is under
way in Lesotho to develop broader public-private coordi-
nation on skill development, it is too early to comment on
its success. Among the commendable initiatives in this
direction has been the development of an interactive data-
base that would collate information on leading global and
regional experts in specific areas. Depending on the inter-
est from the private sector, selected experts would be
invited to develop training modules that would feed into
this interactive database. Eventually, private sector partici-
pants would be expected to use those modules to commis-
sion training at their own expense. The success of such an
initiative will depend on the degree of collaboration among
various stakeholders. A breakthrough in this area will be
necessary for Lesotho to climb up the global apparel qual-
ity value chain.
Transport, logistics, and customs
The time and cost of transport is often an important barrier
to competitiveness in the apparel industry. A smooth trans-
port corridor supported by competent logistical service
providers and punctuated by efficient border crossing mech-
anisms is a prerequisite for any exporting economy. Thanks
to the critical mass of apparel producers that regularly use
the corridor between Maseru and Johannesburg, Lesotho has
made enormous strides in developing systems for transport,
customs, and logistics.
The road infrastructure in Lesotho is substantially devel-
oped and is well connected to the extensive road networks
of South Africa. Trucking services are provided by private
operators that cater to a wide range of apparel producers.
Although there is only one railhead in Lesotho, it is located
in Maseru, just across the river border with South Africa.
Railway lines, station buildings, and handling facilities in
228 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS
Lesotho are owned and operated by Spoornet, the South
African rail service.
Customs processes in Lesotho have significantly improved
since AGOA took effect. The government has put in place a
rigorous training and regulation system among the customs
staff to meet the requirements of exports under the AGOA
visa scheme and now offers duty rebates on fabrics imported
in the production of apparel exports. Further, a duty credit
scheme was put in place for exporters of manufactured goods
that use raw materials sourced from Southern African Cus-
toms Union countries. For certain factories, customs requires
closer inspection of waste and offcuts disposal before rebates
on imported materials are issued, though such inspections
are expedited after factories establish a good track record or
after in-factory inspections are streamlined with other regu-
latory measures. Such multiple inspection requirements
necessitate an efficient, well-trained customs team to ensure
smooth implementation.
LESSONS FROM LESOTHO’S APPAREL
INDUSTRY EXPERIENCE
Integration into global supply chains has been an important
avenue for growth and competitiveness for Lesotho’s apparel
industry. Geographical advantages and disadvantages—of
weather, location, and connectivity—cannot be ignored, but
discovery and pursuit of new opportunities through correct
formulation of policies and development of effective public-
private partnerships has been the key channel through
which Lesotho has achieved apparel industry success. Sev-
eral key features of this success story can be identified.
First, the government of Lesotho was successful in
attracting foreign investors through its policy of construct-
ing industrial zones and serviced factory shells. Throughout
this process, the government has remained cognizant of the
geographical implications of such facilities, ensuring that
they are not far from the public goods necessary for indus-
trialization. Construction and effective management of the
factory shells compensated for inherent inefficiencies in the
land market in the form of prohibitions on private owner-
ship and sales-based transactions of land.
Second, when AGOA-related opportunities became
available, Lesotho made stronger, clearer efforts to attract
FDI than other countries in Southern Africa. Lesotho not
only moved quickly to administer AGOA in country but
also offered generous tax and financial incentives to early
investors to help develop a critical mass of apparel manu-
facturers that then made way for service providers and reg-
ulatory bodies to join in.
 Third, Lesotho’s export and investment promotion
efforts would not have been effective without the LNDC,
which pursued promotion measures through appropriate
information dissemination models and targeted interest
solicitation. Once a critical mass of apparel investors came
to Lesotho, the LNDC developed an effective one-stop shop
for business development, firm registration, financing, and
state-industry coordination on issues such as tax manage-
ment and access to state industrialization incentives.
Fourth, when labor disputes arose around the industrial
transition, the government of Lesotho quickly stepped in to
mediate among multiple stakeholders who were nurturing
mutual antagonism.
And fifth, although it is too early to make informed ver-
dicts, the government of Lesotho has acknowledged its past
mistakes on skill development and is developing a coordi-
nated skill development initiative in conjunction with the
private sector. If successful, the initiative will compensate
Lesotho for its competitive disadvantage in labor costs.
NOTES
1. As per the definition as used in AGOA, a least developed
country in Sub-Saharan Africa is a country with gross
national product (GNP) per capita of less than $1,500 in 1998.
This provision was renewed in 2006 for the period 2007–12.
2. The Multifibre Arrangement governed textile and cloth-
ing quotas between 1974 and the end of 1994. Upon forma-
tion of the World Trade Organization in 1995, textile
exports became governed by the Agreement on Textiles and
Clothing.
BIBLIOGRAPHY
Edwards, L., and R. Lawrence. 2009. “Lesotho’s Export Per-
formance: An African Successful Story?” National Bureau
for Economic Research, Cambridge, MA.
CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS
Eifert, B., A. Gelb, and V. Ramachandran. 2005. “Business
Environment and Comparative Advantage in Africa:
Evidence from the Investment Climate Data.” Working
Paper, Center for Global Development, Washington,
DC.
DFID. 2006. “Lesotho Garment Industry Subsector Study”
conducted for the government of Lesotho.
EIU (Economist Intelligence Unit). 2007. “Country Profile:
Mozambique.” EIU, London.
FIAS (Foreign Investment and Advisory Services). 2006.
“Market Diversification of the Lesotho Garment Indus-
try.” Discussion draft, Innernational Finance Corpora-
tion and World Bank, Washington, DC.
Global Development Solutions. 2005. “Value Chain Analysis
for Strategic Sectors in Mozambique.” Reston, VA.
Horton, C. 1996. “Capital Flows from South Africa into
Neighboring Countries: The Case of Textiles and Cloth-
ing.” Naledi papers. National Labour and Economic
Development Institute, Johannesburg.
Kaplinsky, R., D. McCormick, and M. Morris. 2006. “Dan-
gling by a Thread: How Sharp Are the Chinese Scissors?”
Institute of Development Studies, University of Sussex,
Brighton, United Kingdom.
Minor, P. 2005. “Survey of US Apparel Buyers: Sourching
from Sub-Saharan Africa in the Post-Quota MFA.”
Unpublished paper, Nathan Associates Inc. Arlington,
VA.
———. 2007. “Developing a Cotton Textile and Apparel
Value Chain in Mozambique.” Nathan Associates Inc,
Arlington, VA.
Morris, M., and L. Sedowski. 2006. “Report on Government
Responses to New Post-MFA Realities in Lesotho.”
PRISM Working Paper, Centre for Social Science
Research, University of Cape Town.
World Bank. 2006. “Lesotho Country Economic Memoran-
dum: Growth and Employment Options Study.” World
Bank, Washington, DC.
229
 CHAPTER 14
The Success of Tourism in Rwanda:
Gorillas and More
Hannah Nielsen and Anna Spenceley
wanda is well known for its mountain gorillas. First
R
brought to international attention by the conserva-
tion efforts of Dian Fossey in the 1960s and 1970s,
Rwanda’s gorillas have been featured in numerous documen-
taries and have been visited by well-known figures such as Bill
Gates, Natalie Portman, and Ted Turner, all of whom have
participated in Rwanda’s annual gorilla-naming ceremony.
Rwanda and Uganda are currently the only two coun-
tries in the world where mountain gorillas can be visited
safely, and the number of tourists visiting the Volcanoes
National Park (VNP) has increased dramatically since the
end of the war. Rwanda also views gorilla tourism as a valu-
able conservation tool, and as such enforces strict rules for
the habituation of, and trekking with, gorilla families.
Tourists are willing to pay high fees for a limited number of
permits, which are usually sold out. Revenues from gorilla
tourism provide funds to national parks and facilitate con-
servation activities.
Although Rwanda is known for its violent past, interna-
tional perception of the country is shifting. As of 2010
Rwanda is considered one of the safest destinations in East
Africa. This rebranding goes hand in hand with the market-
ing of the country and, in particular, the mountain gorillas.
The revival of gorilla tourism demonstrates that with the
right strategy, a postconflict country can successfully focus
Hannah Nielsen is an economist in the Poverty Reduction and Economic Management Network of the Africa Region of the World Bank. Anna Spenceley,
PhD, is an independent consultant based in South Africa, and formerly a senior tourism advisor at SNV in Rwanda.
on high-end tourism while maintaining conservation and
contributing to poverty reduction through the involvement
of communities.
Besides the VNP, Rwanda has two other national parks
that offer a range of wildlife and biodiversity. Furthermore,
the country has been particularly successful in attracting
large numbers of business and conference travelers, mainly
from the Democratic Republic of Congo and other neigh-
boring countries of the East African Community (EAC).1
This success is evidenced by the large increase in the number
of hotel rooms, restaurants, and the planned construction of
a convention center. Local and foreign direct investments
have been substantial, accounting for 16 and 20 percent of
total local and foreign direct investment, respectively, over
the last 10 years. In terms of export revenue, tourism already
outperforms coffee and tea by a wide margin.
Several key characteristics have contributed to the suc-
cessful revival of the tourism sector in Rwanda. First and
foremost, the government has shown a clear commitment
to the development of tourism and has established Rwanda
as a safe destination in the region. The early development
of a strategy and policy demonstrated this commitment.
Furthermore, the government involved the private sector
from the start and has implemented policies that enhanced
the business environment and promoted private sector
231
 investment in tourism, thereby marketing Rwanda as a des-
tination. The business environment has improved
markedly in recent years, promoting private sector involve-
ment in tourism. In addition, Rwanda has always seen
tourism as an instrument to reduce poverty, for example by
directly involving local communities.
THE SUCCESS OF GORILLA TOURISM
Background: How did Rwanda start to develop
gorilla tourism?
The Virunga mountain gorilla (Gorilla beringei beringei) is
a highly endangered African ape subspecies, with a total
estimated population of 380, that exists only in the
Virunga Conservation Area encompassing Rwanda, the
Democratic Republic of Congo, and Uganda(figure 14.1).
The distribution of the Virunga mountain gorillas is lim-
ited to an approximate area of 447 square kilometers,
which encompasses the Mgahinga Gorilla National Park in
Uganda, the VNP in Rwanda, and the Mikeno sector of the
Parc National des Virunga of the Democratic Republic of
Congo (Gray et al. 2005). The VNP consists of about
160 square kilometers of montane forest. Until Rwanda’s
independence in 1962, the VNP was part of Africa’s first
national park, the Parc National Albert, which was created
in 1925 with the intention of protecting the great apes
(ORTPN 2004).
Tours to view wild mountain gorilla groups have been
organized since 1955 (Butynaski and Kalina 1997), with
Figure 14.1 Area of Distribution of the Mountain Gorillas
DEM. REP. OF CONGO
Visoke,
3711 m.
Karisimbi,
4507 m.
PARC NATIONAL
DES VOLCANS
Source: Bush and Fawcett 2007.
232 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
the first attempts at habituation for this purpose occur-
ring as early as 1966 (Murnyak 1981). These early tourism
programs displayed an almost complete lack of structure
and control. The focus was on revenue rather than con-
servation, and there are many anecdotal reports of large
groups of tourists visiting groups of nonhabituated or
semihabituated gorillas (Fawcett, Hodgkinson, and
Mehlman 2004).
In 1979 the Virunga region’s first official mountain
gorilla tourism program was launched by Bill Webber and
Amy Vedder with funding from the African Wildlife Foun-
dation, World Wide Fund for Nature, and Fauna and Flora
International (Bush 2009). It was one part of the three-
part approach of the Mountain Gorilla Project, which also
included antipoaching and education programs. The
gorilla tourism program had a dual purpose: providing the
Rwandan government and park authorities an incentive to
conserve the VNP and the animals within it from the
threat of proposed conversion of 5,000 hectares of the
VNP for agricultural purposes; and generating local
employment and tourism-related revenue (Weber 1982,
1985; Vedder and Weber 1990). The program subsequently
evolved into what is now the International Gorilla Conser-
vation Program, still organized as a coalition of the three
agencies (Bush 2009).
Two wild groups of gorillas were initially habituated for
tourism visitation purposes, with strictly enforced limits on
the number of visitors and length of visits (box 14.1). The
combination of quality control and international interest in
IBRD 38552
MAY 2011
PARC NATIONAL UGANDA
DES VOLCANS
Gahinga,
3474 m. Muhavura,
ORTPN H.Q. 4127 m.
Kinigi
Ruhengeri
R WA N D A
0 5 10 Kilometers
 Box 14.1 Current Rules Controlling Gorilla Tourism in Rwanda
A number of rules designed to protect both gorillas and
tourists have been established, as follows:
■ Maintenance of a distance of 7 meters between the
tourists and the gorillas
■ A maximum of eight tourists per visit
■ A limit of one tourist group per day to each gorilla
group
■ A limit of one hour per visit
■ Tourists who are visibly unwell or declare themselves
to be ill may not visit
As of 1989 the number of tourists allowed in a single
visit was increased to eight people (six people for
smaller gorilla groups). In 1999 the required separation
distance between tourists and gorillas (to reduce the
risk of disease transmission) was increased from 5 to
7 meters. Other rules have been added over time:
■ Minimum age of 15 years for tourists
■ No flash photography
Source: Fawcett, Hodgkinson, and Mehlman 2004; Homesy 1999.
Dian Fossey’s highly publicized gorilla studies resulted in
steadily increased visitation throughout the 1980s, peaking
around 6,900 in 1989 (ORTPN 2008b). By the mid-1980s
local attitudes toward and political support for conservation
increased significantly as a direct result of this program
(Weber 1987). Stimulated by the attraction of gorilla
tourism, Rwanda received almost 22,000 visits to its three
national parks in 1990 (Bush, Hanley, and Colombo 2008),
before conflict brought tourism to a halt.
Since the VNP reopened in 1999, visitation has
rebounded from 417 visits that year to nearly 20,000 visits
in 2008 (of which 17,000 were to see the mountain gorillas)
(ORTPN 2008b). Gorilla visitation shows some amount of
seasonality, with a peak in the number of permits sold
between June and September (figure 14.2).
Factors contributing to the success of
gorilla tourism
A number of factors have contributed to the success of
gorilla tourism in Rwanda. A prerequisite is the relative ease
of habituating mountain gorillas to the presence of humans,
facilitated by the temperate climate and benign habitat. The
only other country where mountain gorillas can currently
CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
■ Tourists must remain together in a tight group
■ No loud noises or pointing
■ Eating, drinking, and smoking are not permitted
within 200 meters of the gorillas
■ Tourists must turn away and cover their mouths
when coughing and sneezing
■ Human feces must be buried in a hole of a minimum
depth of 30 centimeters
■ No trash may be deposited in the park
■ Tourists are not allowed to clear away vegetation to
get a better view
These rules (adapted from Litchfield 1997) were
designed and set to minimize behavioral disturbance
and disease transmission to the gorillas from tourists.
Although the welfare of the gorillas has always been the
primary concern, the majority of these regulations were
created based on expert opinions rather than specific
research findings.
be visited safely is Uganda. With a broad client base2 and a
limited number of permits (around 17,000 per year),
demand is higher than availability of permits. The accessi-
bility of the gorillas is another advantage. Because of
Rwanda’s small size, tourists can reach the gorillas in two
hours from Kigali; by comparison, it takes six hours to reach
the gorillas from Kampala in Uganda. In addition, the con-
dition of infrastructure in Rwanda, especially roads, is rela-
tively good compared with that of its peers.
Besides viable tourism assets and relatively good infra-
structure, Rwanda has shown a strong commitment to pro-
moting the tourism sector. It has developed a clear tourism
strategy, marketed the destination Rwanda successfully,
involved the private sector in the policy dialogue, and gen-
erally improved the country’s business environment.
Overall strategy and vision. Between 1994 and 2001
the Rwandan government worked to establish a tourism-
friendly environment. The first meetings with the private
sector on the development of the tourism sector were held
in 1999. From 2000 onward Rwanda participated in major
tourism fairs, and in late 2001 the Tourism Working Group,
which included the public and the private sectors, was
established. The Rwanda Tourism Strategy was developed
233
 Figure 14.2 Gorilla Trekking in Volcanoes National Park
a. Park activities in VNP, by product, 2008
Golden 2,000
monkey 1,800
7% 1,600
1,400
Gorilla 1,200
85% Mountain
climbing 1,000
5% 800
600
Other
400
3%
200
0
Fe ry
ua
ua
Ap
ar
Jan
br
Source: ORTPN 2008b.
and approved by the Cabinet in 2002. A revised Rwanda
Tourism Strategy (“Sustaining the Momentum”) was elabo-
rated in 2007. A National Tourism Policy was put in place in
2006, a revision of which is currently under way. With sup-
port from the United Nations World Tourism Organization,
the government of Rwanda has also prepared a 10-year Sus-
tainable Tourism Development Master Plan (Republic of
Rwanda 2002, 2007b, 2009a, and 2009b).
Rwanda’s overall strategic vision is to focus on high-end
ecotourism rather than mass tourism. In the first Rwanda
Tourism Strategy, three core market segments were identi-
fied: ecotravelers, explorers, and business travelers. The tar-
gets set in that document were soon surpassed, however,
mainly through the success of the gorilla product. The
revised strategy in 2007 identified the primates as Rwanda’s
unique selling proposition but recognized the need to diver-
sify the tourism sector and identified international confer-
ences and birding as two additional core segments. Tourism
receipts are already higher than the targets set for 2012 in
the revised strategy. In the latest tourism policy, objectives
are set within the framework of other national strategic
documents, such as the Vision 2020 and the Economic
Development and Poverty Reduction Strategy.
The Sustainable Tourism Development Master Plan for
Rwanda consolidates previous strategies and policies, gives
clear and detailed recommendations, and sets ambitious
targets. Tourist arrivals to Rwanda are projected to increase
from about 980,000 in 2008 to more than 2 million in 2020,
with an expected increase in foreign exchange earnings
from about $200 million to more than $600 million. Sepa-
rate strategies are being developed for MICE (meetings,
234 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
b. Gorilla trekking permits by month, 2008
ry
ch
ril
ay
ne
ly
st
ov er
r
be
be
be
Ju
gu
ob
Ju
em
em
em
Au
ct
pt
ec
O
Se
D
N
incentives, conferences, and exhibitions) tourism and for
birding activities (OTF Group 2008a, 2008b).
Although the different strategies have not yet been
implemented completely, the Rwandan government has
consistently demonstrated strong commitment to the exe-
cution of tourism-related reforms and to the overall
improvement of the performance of the sector. Despite the
limited number of staff in the Rwanda Office of Tourism
and National Parks (ORPTN), its leadership has effectively
led the advancement of reforms. Further, Rwanda has
learned from the experience of other countries. Study tours
have been undertaken to Kenya and Mauritius to learn from
the tourism development strategies of those countries.
Marketing. Following the passing of formal tourism plans
in Rwanda, a national campaign was launched to improve
the image of tourism in the country. (The word for tourism
in Kinyarwanda, the local language, means “wandering
around aimlessly” and has therefore a negative connotation.)
A media campaign was launched to sensitize the population
and convey that the country can benefit from tourists and
should therefore welcome foreigners. Simultaneously,
Rwanda has worked to improve its image on an interna-
tional basis. In the late 1990s international perceptions of
Rwanda were primarily associated with the genocide. Even
still in 2002 market research conducted in neighboring
countries showed that more than half of international visi-
tors believed that Rwanda was an unsafe destination
(Grosspietsch 2006). Despite concerns about the safety situ-
ation, however, surveys carried out in 2003 showed that
the satisfaction level of visitors in terms of safety and
stability was very high after they had visited the country
(Grosspietsch 2006).
Specific public relations and marketing efforts pursued
by Rwanda include contracting international public rela-
tions and marketing agencies in the United Kingdom and
United States and launching a new Web site in 2003. In
addition, Rwanda has been featured extensively in docu-
mentaries on international television channels and has
received positive coverage in more than 350 credible inter-
national press publications, as well as in major travel
guides. Rwanda has also represented itself well at major
tourism fairs since 2000, earning first prize for the best
African stand at the International Tourism Bourse in Berlin
for three consecutive years, 2007–09, and at the World
Travel Market in London in 2009. ORTPN’s financing of
travel fees for several Rwandan tour operators to these
trade fairs has contributed to cooperation between the gov-
ernment and the private sector on issues related to tourism.
To foster the interest of tour operators and travel agents in
Rwanda, the government has also organized familiarization
tours for international investors and tour operators, during
which the Minister of Commerce and ORTPN have
received the delegates to demonstrate the importance of
their visit.
Rwanda’s annual gorilla naming ceremony (Kwita Izina),
launched in 2005, during which mountain gorillas born in
the previous 12 months are named, has attracted a number
of international celebrities. The baby gorillas have been
named, among others, by the president of Rwanda and his
wife, ambassadors, Hollywood stars, international conserva-
tionists, and performing artists. The ceremonies provide a
good platform to promote Rwanda as a destination and the
need for efforts to protect gorillas and conserve their habi-
tat. The ceremony is now accompanied by several other
events, including a cross-country cycling tour and a conser-
vation conference. Thanks partly to the awareness of the
need to protect the gorillas that the gorilla naming cere-
mony and gorilla tourism in general have brought about,
poaching has been significantly reduced and the number of
gorillas has increased steadily.
Improved business environment and involvement of
the private sector. Initially, promotion of Rwanda’s
tourism sector was almost entirely driven and implemented
by the government. The private sector lacked the capacity
and funding and was not well organized. The government,
however, made efforts to involve the private sector from the
start with the long-term objective that the private sector
would take over as the driving force for encouraging
CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
tourism in Rwanda. As a result, there is now a strong public-
private dialogue surrounding tourism in Rwanda, and a
tourism working group composed of private and public
stakeholders in the tourism industry is now in place. The
private sector is consulted in the development of new
policies and strategies, such as the Sustainable Tourism
Development Master Plan. In addition, the private sector
federation, of which the tourism chamber is a member, is
consulted before new strategies and laws are adopted.
Rwanda has implemented a number of market-based
reforms to strengthen the role of the private sector in
tourism. Several important laws and codes have been
revised, including the investment code, company law, secure
transactions law, labor law, and insolvency law. The new
insolvency law facilitates the access to finance, allowing mov-
ables, such as livestock, to be used as guarantee. Customs pro-
cedures are also being simplified. A pilot, one-stop-window
was successfully launched at one border crossing, and plans
are in the works to replicate it at other border posts.
Rwanda’s business environment has also improved sub-
stantially in recent years. A one-stop-window has been
introduced to register a business, and the administrative
costs of registering a business have been lowered. It is now
possible to register a business within one day for a flat fee of
RF 25,000 ($43). Rwanda’s success in this area has been doc-
umented by a substantial improvement in the World Bank’s
Doing Business indicators: Rwanda was named the top per-
former in 2009. Rwanda outperforms all other countries in
the EAC in the rankings and has shown a strong commit-
ment to further improving private sector conditions, partic-
ularly in the tourism sector.
A number of tourism sector–related incentives are
offered to investors. According to the investment code, tax
exemptions are granted to investors who invest $100,000 or
more in a facility. Airplanes imported to transport tourists
are tax exempt, and specialized vehicles such as hotel shut-
tles are exempt from import and excise duties. An investor
in the tourism and hotel industry is also exempt from pay-
ment of import duties on equipment such as bedroom fit-
tings, swimming pools, and outdoor leisure equipment.
BENEFITS OF GORILLA TOURISM
Implications for communities
Bush, Hanley, and Colombo (2008) note that the Virunga
mountain gorilla represents an isolated island population in
an upland area surrounded by a sea of humanity at some of
the highest densities found on the African continent (in some
areas of Rwanda, population density reaches 820 people per
235
 square kilometer), and much of the land surrounding areas
where gorillas live is inhabited by extremely poor people who
derive their incomes from agricultural activities. Gorillas are
severely threatened by anthropogenic disturbances such as
agricultural conversion and illegal extraction of resources (for
example, snare setting for smaller mammals that entrap
young gorillas). While gorillas are no longer hunted for their
meat in the Virunga range, they are the focus of illegal animal
trafficking. Members of gorilla groups are killed and
wounded in an effort to trap infants for the black market,
sometimes leading to the disintegration of groups. This hunt-
ing pressure currently represents the greatest threat to the
survival of the mountain gorillas and the integrity of their
habitat. Illegal hunting is mainly motivated by meeting sub-
sistence needs for the poorest people around the VNP
(Plumptre et al. 2004).
To address local welfare needs to mitigate some of these
poverty-related conservation threats, a key focus of contem-
porary conservation strategies is on local communities
(Hulme and Murphee 2001). Combining conservation with
local development through integrated conservation and devel-
opment projects is now a standard approach in many devel-
oping countries (Barrett and Arcese 1995). Rwandan commu-
nities are involved in gorilla tourism in the following ways:
■ Creation of a department for community conservation
to work on local education and social infrastructure
projects (Uwingeli 2009).
■ Revenue sharing: Since 2005 ORTPN (which was
absorbed into the Rwanda Development Board, or RDB,
in early 2009), with the support of the government, has
Box 14.2 Process for Disbursing Community Funds
Five percent of tourism revenues from the protected
areas in Rwanda are put into a fund for community
projects in administrative sectors that neighbor
national parks. The Rwanda Development Board
issues calls for proposals, and a project selection
process is completed at the sector and district levels.
Selection criteria include positive impacts on local
communities and on conservation of biodiversity in
protected areas. Areas that register a large number of
cases of conflict between protected areas and the com-
munity, according to the results of a ranger-based
monitoring system, have preferential access to funds,
as do areas that are located close to the protected areas.
Source: Tusabe and Habyalimana 2010.
236 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
overseen a revenue-sharing scheme whereby 5 percent
of tourism revenues from VNP fees are injected into
local community projects around the national park to
ensure that the local people feel some ownership of the
parks (box 14.2). Although it is not known what pro-
portion of the budget of local councils is represented by
the shared revenue, it is clear that local governments
must be actively involved in selecting local projects to
finance.
■ Employment opportunities are offered through
national parks: guides, trackers, and antipoaching agents,
for example. Some of the private tour operators also offer
community-based tourism activities, such as stays with
local families, village walks, banana beer production, and
even volunteer opportunities in local communities.
Disbursement of funds to communities. Since 2005
nearly $428,248 has been directly invested in community
projects and used to empower communities. The total
amount, however, equates to an investment of only $1.45
per person since the program’s inception, or an average of
$0.36 per person per year. Projects for which funds have
been used include education, environmental protection
(tree planting, soil erosion control, and fencing in protected
areas to limit access by poachers), food security, basic infra-
structure, and water and sanitation (figure 14.3). Specific
community projects have included construction of schools,
water tanks, and hospitals; basket weaving; establishment of
culture centers; potato farming; tree planting; bee-keeping;
milk cooler construction; goat rearing; and mushroom and
pepper farming. Education projects have received the most
Sustainability of projects (gauged through the eco-
nomical, social, and environmental indicators stated
in the proposal and their likelihood of being achieved)
and the proportion of community contribution are
also considered.
Once projects are selected, contracts are signed with
the district authority and the community. Contracts
cover lengths of time that depend on the project com-
plexity and can vary from 1 month to 15 months. The
community is often grouped into cooperatives or
direct-specific target groups when their ownership and
level of organization guarantee effective implementa-
tion of the project.
Figure 14.3 Funds Disbursed to Community Projects around the VNP, by Sector, 2005–08

Water $8,772

Infrastructure $15,789
Types of projects
Food security $40,350

Environment protection $42,105

Education $321,232

$0

00

00

00

00

00

00
00

,0

,0

,0

,0

,0

,0
0,

00

50

00

50

00

50
$5

$1

$1

$2

$2

$3

$3
Value in US$
Source: Télesphore 2009.

Figure 14.4 Funds Disbursed to Community Projects

funds because of the high priority education is given within
around the VNP, 2005–08
the sectors in the Musanze district, near the VNP. The
annual amount disbursed is directly correlated to tourism 140,000
revenues collected in the previous year. The amount of
funds disbursed to communities in each year between 2005 120,000
and 2008 is shown in figure 14.4. 100,000
By 2008 seven districts bordering parks in Rwanda with
U.S. dollars

a population of almost 300,000 people had been reached by
the community project financing scheme (Bush 2009).
Although no formal study has yet been carried out to assess
the impact of the scheme on the livelihoods of people living
near the VNP, the RDB and local authorities indicate that
the scheme has contributed to an increased awareness of
tourism benefits to the community and to the need to pro-
tect biodiversity in the VNP (Spenceley et al. 2010).
One specific project that has benefited from community
fund financing is the high-end Sabyinyo Silverback Lodge.
The eight-room lodge located at the periphery of the VNP
is a joint venture of the local Kinigi and Nyange communi-
ties (represented by the Sabyinyo Community Livelihoods
Association, or SACOLA); the private sector (Governors
Camps Ltd); international NGOs, in particular, the Interna-
tional Gorilla Conservation Program (IGCP) and the
African Wildlife Foundation (AWF); and the RDB. Planning
for the lodge began in 2004, and the first tourists began to
arrive in August 2007 (Makambo 2009). Some initial fund-
ing was obtained from the U.S. Agency for International
Development (Verdugo 2009). The joint venture agreement
includes a 15-year lease agreement between Governors
Camps Ltd and SACOLA. The private sector operator built
80,000
60,000
40,000
20,000
0
2005 2006 2007 2008
Source: Télesphore 2009.
and operates the lodge, and pays the communities a $50
occupied bed night fee and also 7.5 percent of net sales
(Makambo 2009).
This joint venture operation allows people who live
close to the VNP to benefit from tourism in four main
ways: equity in a tourism business, employment opportu-
nities at the lodge, the supply of goods and services, and
dividends from profits. The lodge employs 45 local people,
who receive training and experience in hospitality and
tourism. Local agricultural produce is purchased for use by
the lodge, and there are plans to establish traditional danc-
ing, a cultural center, a community walk, and handicraft
sales at the lodge. In addition, there are plans under the

CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 237
 joint venture to use funds from the lodge to finance infra-
structure, such as roads, in the area (Makambo 2009).
Weaknesses of the joint venture, however, include poor
leadership and weak governance of SACOLA and the
reliance on the private sector operator to generate revenue
from tourists (Makambo 2009).
Social benefits from gorilla tourism. Since the
tourism revenue-sharing scheme was initiated in 2005, a
number of direct and indirect projects with social benefits
of gorilla tourism for local communities living near the
VNP have been implemented (Uwingeli 2009):
■ Schools: Ten schools have been constructed, with 56
classrooms and an average of 65 pupils per classroom per
rotation (morning and afternoon). The school construc-
tion has reduced the distance traveled by children to the
nearest schools, allowing them to spend more time on
their studies after school.
■ Water tanks: Thirty-two water tanks have been con-
structed. These provide 20 liters of water per person per
day, and at least 1,250 people are served by each tank.
■ Income-generating activities: Ten community associa-
tions have been supported directly through the revenue-
sharing scheme, and a number of other projects, such as
bee-keeping and basket weaving, have been imple-
mented. In all of these projects, the focus has been on
training for income-generating activities.
■ New partnerships in conservation and community
development brought to the construction of the
Sabyinyo community lodge, which is owned by
SACOLA, but managed by a specialized ecolodge com-
pany. At least 3,000 households are members of SACOLA
and benefit from the agreement with the managing com-
pany to pay SACOLA bed night fees and a percentage of
monthly net income.
Employment from gorilla tourism. The VNP employs
at least 180 people, who work as guides, gorilla group track-
ers (for both tourism and research groups), and antipoach-
ing teams deployed in the five protected sectors of the VNP
(Uwingeli 2009). In addition, an estimated 800 community
members around the VNP are involved in day-to-day man-
agement activities and benefit from opportunistic and tem-
porary employment and the revenue-sharing scheme. VNP
management has helped to form two umbrella associations:
one for conservation activities in the VNP (Amizero, or
hope) and another for community development activities
(Iby’Iwacu). Several hundred volunteers work with the two
238 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
organizations as crop rangers, conservationists, porters, and
community awareness representatives.
The Iby’Iwacu cultural village has been developed in col-
laboration between a private sector tour operator, Rwanda
Ecotours, and a group of former poachers living near the
VNP. As a result of an academic research project, a partici-
patory process began in 2005 to transform the livelihoods of
poachers toward farming, and then tourism. Part of the
concept was to benefit conservation, by providing alterna-
tive livelihood opportunities to illegal hunting of buffalo
and other wildlife in the VNP. Meetings with poachers were
held to gain their trust and insights, and study tours were
undertaken to raise understanding of cultural tourism
products elsewhere.
Interviews were undertaken with tourists to establish
what they would be willing to pay for a cultural experience.
Community members engaged in the design, construction,
and operation of the cultural village. Local architectural
techniques were used in the construction along with local
materials such as thatching grass and wood. Tourists can
stop at the Iby’Iwacu and experience local attire, practice
traditional fire-making techniques, archery, drumming, and
dancing; visit a traditional healer; and prepare and eat tra-
ditional food. The village generates around $14,000 per year,
and community representatives identify projects in the
community that they can finance (Sabuhoro 2009).
Challenges. Studies show that tourists to Rwanda do not
have a particular willingness to pay for community bene-
fits. Bush, Hanley, and Colombo (2008), for example, find
that the percentage of VNP revenues used to enhance local
community development does not have a significant effect
on tourism demand. Bush, Hanley, and Colombo also note
that these findings do not imply that tourists are unwilling
to take part in community-based tourism—rather, that
they are not willing to sacrifice other immediate benefits of
the trekking experience relative to increases in permit
prices that were dedicated to revenue sharing. This repre-
sents an important departure from common ecotourism
principles about social benefits. Bush (2009) suggests that
tourists need to be better educated about the human
dimension of conservation to emphasize the conceptual
link between the needs of the local populations and biodi-
versity conservation.
Furthermore, Rutagarama and Martin (2006) state that
there is something of a catch-22 in relation to community
conservation in Rwanda. On one hand, the empowerment
of local partners will be constrained when appropriate
powers are not devolved to them. On the other hand, it is
impossible to impose powers on those who feel neither
capable nor inclined to exercise them. Rutagarama and
Martin (2006) therefore suggest the need for a flexible
framework that enables capacity and power to coevolve in
locally appropriate ways. Developing the assets that part-
ners need to maximize their opportunities for entering
productive partnerships should be a fundamental part of
plans to widen (and deepen) local participation in tourism.
In some countries this has been done through a hybrid
approach, and by retrofitting community-based tourism
enterprises into joint-venture partnerships with private
sector operators. Although the transaction costs, such as
the time needed to negotiate deals, can be high, these part-
nerships can provide a win-win for the private sector and
community members with appropriate agreements (exam-
ples are Damaraland in Namibia, Rocktail Bay and Phinda
Reserve in South Africa, and Covane Fishing and Safari
Lodge in Mozambique). In these situations the capacity of
community members is built over time through partner-
ships with businesses that understand the tourism sector,
how to operate a business, and how to establish market
linkages.
Implications for conservation
Mountain gorilla tourism in Rwanda has long been viewed
as a valuable conservation tool. An economic incentive to
conserve the mountain gorilla is provided by international
tourists paying relatively large sums of money to spend a
short amount of time with the gorillas. Since its conception,
organized gorilla tourism has provided funds to VNP
authorities to assist with conservation activities. Nature-
based tourism has thus been enthusiastically accepted and
supported by governments, conservationists, and tourists
alike and, in Rwanda, has been acknowledged as playing a
crucial role in the success of mountain gorilla conservation
in the VNP (Bush, Hanley, and Colombo 2008).
The number of mountain gorillas left in the world is esti-
mated to be approximately 700. As of 2003, 380 gorillas
lived in the Virunga volcanoes range, while 320 lived in
Bwindi Impenetrable National Park in Uganda as of 2006
(Uwingeli 2009). Research indicates that the gorilla popula-
tions in areas frequented by tourists are increasing, with an
overall growth in the population of 1.1 percent between
1989 and 2003 (Fawcett 2009) (box 14.3). Gorilla groups in
the Democratic Republic of Congo are doing less well in
terms of population growth, which may be due to the lower
presence of patrols, researchers, and tourist visits there com-
pared with Rwanda and Uganda (Fawcett 2009).
CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
Box 14.3 Growth in Gorilla Conservation Efforts
Mountain gorillas in the Virunga region of Rwanda,
Uganda, and the Democratic Republic of Congo were
censused five times between 1970 and 1989, when the
population was estimated at 324 gorillas. War and
political unrest in the region since 1990 meant that no
census was conducted over the next decade, and in
2000, the observation of 32 groups provided an esti-
mate of between 359 and 395 gorillas. This represents
a 0.9 percent to 1.8 percent annual growth rate over
10 years and a 1.0 percent to 1.3 percent annual
growth rate between 1972 and 2000.
Source: Kalpers et al. 2003.
Biological research on mountain gorillas by teams of
researchers in the Karisoke Research Centre in Rwanda pro-
vides important information on trends in sound tourism
management. Of the estimated 380 gorillas in the Virunga
Volcanoes Range, at least 260 are habituated and regularly
monitored in Rwanda (Uwingeli 2009): These gorillas are
checked on a daily basis; health reports are shared, and
actions taken when necessary. The habitat is patrolled daily
to detect illegal activities and discourage attempts to set
snares. The use of information technology allows the habi-
tat of the mountain gorillas to be mapped—in particular, to
show where gorilla groups are moving and what illegal
activities are occurring and to plan ranger patrol activities
accordingly.
Over the years, the Rwandan government has become
more supportive of gorilla conservation, including allocat-
ing more land around the VNP for cultivation to reduce
pressure on the park for agriculture and natural resource
use (Uwingeli 2009). A consultation exercise is currently
under way to assess the feasibility of a VNP expansion pro-
gram (Bush 2009). Although the number of snares found in
the VNP have increased over time (Fawcett 2009), some for-
mer poachers have begun working on conservation efforts,
and there is even an “ex-poachers association” consisting of
about 400 local community members, who patrol with
ORTPN staff and also help with local education, collecting
information, and addressing human-wildlife conflict (for
example, crop raiding) (Uwingeli 2009).
Although no systematic method currently exists for regis-
tering or training guides in Rwanda, VNP staff have bene-
fited from capacity-building programs by the RDB (and
previously, by ORTPN) and their partners in guiding, gorilla
239
 health monitoring, and general biodiversity conservation.
Also, with the support of IGCP and the Karisoke Research
Centre, additional training has been provided for gorilla
trekking guides on how to work with visitors and how to
minimize adverse impacts on the gorillas (Kalpers et al. 2003;
Fawcett 2009).
Revenues from Rwanda’s national parks are primarily
used to fund conservation efforts in the parks and world-
wide tourism marketing activities. Salaries for all of the staff
are paid out of national park fees. VNP park management is
cofunded through the research activities of the Karisoke
Research Centre, which provides basic park management
functions such as monitoring and antipoaching patrols for
gorillas. Further conservation funds are contributed by
NGOs such as the IGCP and CARE. These additional funds
and support have contributed enormously to gorilla conser-
vation successes in Rwanda (Bush 2009).
Willingness to pay for conservation
Research by Bush, Hanley, and Colombo (2008) finds that
gorilla trekking tourists are willing to pay for biodiversity
conservation, for both gorillas and other wildlife. The
authors also find that tourists prefer to be in small
Figure 14.5 Impact of the 2007 Gorilla Permit Price Increase on Visitation
a. Tourist visitation by income group
50
40
30
Percent
20
10
0 0
<$25,000 $25,000 to $45,000 to $65,000 to >85,000
$45,000 $65,000 $85,000
Respondent annual income group
After price rise
Source: Bush and Fawcett 2007.
240 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
groups3 and prefer the length of their trek to be between
one and three hours. These two findings could be inter-
preted as showing that tourists support the ecotourism
principle of minimizing ecological impact, since more
people taking longer trips would increase adverse ecolog-
ical impacts.
However, price increases in gorilla permits appear to
affect tourism visitation, at least in the short term. Research
by Bush and Fawcett (2007) reveals that the price increase in
June 2007 from $375 to $475 had a marked impact on the
demographics of visitors to the mountain gorillas (figure
14.5). In particular, the percentage of visits by people in the
highest income group increased significantly, while visits by
all but one of the other income groups decreased. The
length of the stay also decreased significantly, from a mean
number of 4.2 nights to 3.6 nights.
The proportion of visitors going to the other national
parks in Rwanda was also significantly lower after the price
increase, as was the proportion of visitors to the genocide
memorial and taking the Kigali city tour. The proportion of
visitors participating in alternative activities within the VNP
increased, however: hikes on the Karisimbi and Bisoke vol-
canoes, viewing golden monkeys, visits to Dian Fossey’s
tomb, and nature walks.
b. Length of trip
50
40
30
Percent
20
10
1 2 3 4 5 6 8
Nights stayed in Rwanda
Before price rise
 The reduction in visitors’ length of stay and in the fre-
quency of other activities caused by the price increase in
gorilla tourism has implications for the overall economic
impact of international tourism in Rwanda. However, Bush
(2009) notes that it is probably time to repeat the study and
establish whether the changes in demographics and con-
sumption were temporary or lasting. This study would also
provide guidance for decision making about further price
changes. A study of tourism satisfaction and pricing for
alternative products to ensure value for money is also
needed in order to increase the number of bed nights and
overall trip spending (Bush 2009).
Remaining constraints and lessons learned
There are a number of constraints to the further development
of gorilla tourism in Rwanda:
■ Gorilla tourism can present a threat to gorilla conserva-
tion, affecting, for example, the health and behavior of
gorillas (box 14.4), and needs to be well managed.
Box 14.4 Threats to Gorilla Conservation from Tourism
All six great apes—gorillas Gorilla gorilla and Gorilla
beringei, chimpanzees Pan troglodytes and Pan paniscus,
and orangutans Pongo pygmaeus and Pongo abelii—are
categorized as endangered on the International Union
for Conservation of Nature’s 2000 Red List. Threats to
these species include loss of habitat to settlement, log-
ging, and agriculture; illegal hunting for bushmeat and
traditional medicine; the live ape trade; civil unrest; and
infectious diseases. The great apes are highly susceptible
to many human diseases, some of which can be fatal. If
protective measures are not improved, ape populations
that are frequently in close contact with people will
eventually be affected by the inadvertent transmission
of human diseases. Regulations that protect habituated
apes from the transmission of disease from people are
often poorly enforced. Enforcement of existing regula-
tions governing ape-based tourism, and the risk of dis-
ease transmission between humans and the great apes
minimized (Woodford, Mutynsky, and Karesh 2002).
Tourism might also be changing gorilla behavior in
negative ways. In one research study, Fawcett, Hodgkin-
son, and Mehlman (2004) assessed more than 10 months
CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
■ Although the number of tourist accommodation facili-
ties is growing, the facilities are not sufficient at key
tourism sites, including the VNP. Additionally, the qual-
ity of accommodations is not standardized, and prices
are high relative to accommodations of equivalent qual-
ity in other East African countries.
■ The quality of customer service is inconsistent and gen-
erally poor compared with that in neighboring countries.
Also the focus on high-end consumers implies demand
for higher quality. Bush found that after the increase of
the permit price respondents registered significantly
lower levels of satisfaction with their trekking experience
(Bush 2009).
■ Public and private sector collaboration is improving
through the Joint Action Development Forum and Steer-
ing Committees, though it is still weak.
■ Although the road distance between Kigali and Musanze
is short, the quality of the road could be improved.
■ While regional collaboration is of interest to government
authorities of the Democratic Republic of Congo,
Rwanda, and Uganda, regional instability has made it
difficult to harmonize tourism and conservation in
of behavioral data collected from three gorilla groups
during one-hour observation sessions before, during,
and following tourist visits. Results from these data show
clearly that the current tourism program is having a sig-
nificant impact on gorilla behavior. All three gorilla
groups were found to spend significantly less time feed-
ing and more time moving during tourist visits. In addi-
tion, the frequency of certain aggressive behaviors, many
directed at humans, increased in all three groups during
tourist visits, and the gorillas increased their proximity to
the silverback members (adult male gorillas usually more
than 12 years of age).
Some gorilla behaviors observed during the study
were correlated with the distance maintained between
the gorillas and the tourists and the number of tourists
in the gorilla group. Reducing this impact on gorilla
behavior may be a simple matter of better training
guides to maintain the 7-meter distance rule between
tourists and gorillas. Many of these changes in gorilla
behavior during tourist visits are believed to indicate
higher levels of stress in gorillas (Fawcett, Hodgkinson,
and Mehlman 2004).
241
 accordance with the Virunga Massif Transboundary Plan
(Uwingeli 2009; Mehta and Katee 2005).
■ There is widespread poverty around the VNP and
increasing pressure for more agricultural land by a grow-
ing rural population.
■ Benefits accruing from the VNP should be in relation to
the needs of people living close to the park. Local people
desire individual benefits (such as money), and not only
collective infrastructure that is used by the whole popu-
lation, such as water tanks or clinics (Uwingeli 2009).
Several tourism-related lessons learned at the VNP have
the potential to improve other protected areas in Rwanda:
■ Because of the limited number of available gorilla per-
mits, product diversification and promotion is required
to encourage visitors (particularly repeat visitors) to stay
longer, spend more money, and visit other destinations
in the country.
■ Gorilla conservation needs to be balanced with research
visits and tourism trips to ensure that the health of the
gorillas and the integrity of their habitat are maintained.
■ Conservation efforts focused on the key species is impor-
tant, but the contribution to the habitat/ecosystem con-
servation must also be ensured. Long-term dedication
and partnerships in conservation (research, protection,
and tourism) are essential.
■ Sustainability of gorilla tourism can be achieved only if
regional collaboration is established to conserve trans-
boundary protected areas and cross-border resources.
■ Standardized and high-quality training for guides is
needed, for those working both within and outside pro-
tected areas.
■ A more diverse range of accommodation and restaurant
facilities is required, with higher quality and better value
for money.
TOURISM IN RWANDA: THE BIGGER PICTURE
Besides the mountain gorillas in the VNP, Rwanda has other
excellent tourism assets that create a wider foundation for
the tourism sector. Rwanda has three national parks that
cover about 10 percent of the country’s area, one of which is
the VNP. The Akagera National Park offers a range of wildlife
including elephants, hippopotamuses, giraffes, and zebras.
The Nyungwe Forest National Park has a large tract of
mountain forest and is rich in biodiversity. Guided walks and
chimpanzee tracking is offered. Lake Kivu has recreational
facilities as well, but there is still potential for significant
242 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
tourism development. Rwanda also offers business opportu-
nities, mainly for travelers from the eastern part of the Dem-
ocratic Republic of Congo and other neighboring countries.
Rwanda has also been successful in attracting national,
regional, and international conferences.
Development and structure of the tourism sector
In the 1970s and early 1980s only a small number of inter-
national tourists visited Rwanda. Most tourists visited
Akagera National Park, a government-owned, high-end des-
tination used mostly for hunting. Only a very limited num-
ber of tourists visited the gorillas. Tourism was not a
national priority and was not viewed as a tool to reduce
poverty. The first hotel, the Mille Collines, was built in 1973
and the ORTPN was created in 1974. No tour operators
existed in the 1970s and 1980s, and the sector was domi-
nated by the government, which owned all hotels except the
Mille Collines.
Tourist arrivals started to increase notably in the 1980s
(figure 14.6). Most tourists still visited only Akagera
National Park at that point, but the first official mountain
gorilla tourism program was launched in 1979, leading to
continuously increasing visitor numbers in the VNP, peak-
ing at 6,900 in 1989 (ORTPN 2008b). This trend was
brought to an abrupt end with the outbreak of the war in
mid-1990s, however. Visits to the VNP, which provided
most of the tourism revenue by the mid-1990s, dropped in
1994 due to the genocide and again between 1997 and 1999,
when the VNP had to be closed for some time due to an
insurgency. Visits increased in the years following, however.
In 2008 about 17,000 people visited the VNP to see the
gorillas, a dramatic increase from the late 1980s and an
impressive recovery from only 417 tourists in 1999 after the
reopening of the VNP following the war. Total visits to all
Rwanda’s national parks reached more than 43,000 visitors
in 2008. Today, the majority of visitors to the VNP are for-
eigners, while Akagera National Park is visited by a more
equal mix of Rwandan and foreign residents.
Reliable tourist arrival statistics for Rwanda as a whole are
available only for 2007 onward, when entry cards were intro-
duced. This innovation led to a substantive upward revision
of tourism revenue. The use of statistics on park visits alone
led to a significant underestimation of the total number of
tourists. Most important, the large number of business and
conference tourists had not been taken into account.
Although data are still not collected electronically and the
quality of data is not high, there is little doubt that the recov-
ery of the tourism sector in Rwanda has been successful.
Figure 14.6 Visits to Rwanda’s National Parks, 1974–2008

45,000

40,000

35,000

30,000

Number 25,000

20,000

15,000

10,000

5,000

0
74
76
78
80
82
84
86
88
90
92
94

96
98
00
02
04
06

08
19
19
19
19
19
19
19
19
19
19
19

19
19
20
20
20
20

20
Nyungwe Akagera Volcanoes

Source: ORTPN.

Table 14.1 Tourist Arrivals to Rwanda by Purpose of Visit, 2007–09

Thousands
Visiting friends Conference/
Holiday/vacation and relatives business Transit Other
Share Share Share Share Share
Year Number (%) Number (&) Number (%) Number (%) Number (%) Total
2007 21.5 2.6 332.0 40.2 275.8 33.4 150.1 18.2 47.0 5.7 826.4
2008 59.4 6.1 248.3 25.3 345.9 35.3 307.8 31.4 19.1 1.9 980.6
2009 (Jan - Jun) 21.4 4.9 112.2 25.6 187.9 42.8 83.3 19.0 34.1 7.8 439.0

Source: ORTPN.

A total of 980,577 international arrivals were recorded in
2008, up from 826,374 in 2007 (table 14.1). The main char-
acteristics of international tourists entering Rwanda are:
■ Most visitors came for business and conferences (35 per-
cent in 2008). This had already been indicated in a hotel
market study, which found that 75 percent of all tourists
in the country in 2006 were business travelers (IFC 2007).
■ The share of tourists arriving in Rwanda for vacation is
relatively small, but increased from 3 percent in 2007 to
6 percent in 2008.
■ The large number of international arrivals includes tran-
sit passengers, thereby reducing the overall number of
arrivals that can be counted as tourists entering Rwanda.
The tourist entry cards offer more details about travelers
entering Rwanda. Overall, 88 percent of total international
arrivals came from Africa. The country or region of origin,
however, varies considerably by the purpose of the visit. Most
of the tourists on vacation came from Europe and the United
States (figure 14.7a). The majority of business and conference
travelers came from the Democratic Republic of Congo
and the other EAC member states (figure 14.7b). Most inter-
national tourists visiting friends and relatives in 2008 came
from the EAC and Congo Republic (figure 14.7c). According
to estimates by ORTPN and the Ministry of Economy and
Finance, non-African tourists coming to Rwanda for leisure
or conference and business purposes spent the most money
among the various categories of tourists (table 14.2).
Most leisure tourists visit the region as part of a multi-
country itinerary and do not yet consider Rwanda as a stand-
alone destination. A recent survey of tourists, tour operators,
and accommodation providers in Rwanda (SNV and RDB
2009) finds that the most common length of stay for

CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 243
 Figure 14.7 Country/Region of Origin of Foreign Arrivals to Rwanda, by Purpose of Visit, 2008

a. Holiday/vacation b. Business/conference c. Visiting friends and relatives

USA and
USA and Other Canada
Canada 6% Other
2% 3%
3% Europe
4%

Other Europe
11% EAC 4%
21% EAC
USA and 30%
Canada EAC
24% Congo, Dem.
Rep. 50%
Congo, Dem.
14% Congo, Dem. Rep. Rep.
Europe 57% 41%
30%

Source: ORTPN.

Table 14.2 Average Spending by Tourists in Table 14.3 Growth in Hotel Rooms, Restaurants,
Rwanda, by Purpose of Visit, 2009 Tour Operators, and Travel Agencies in
Average amount spent per visit Rwanda
(US$)
Average annual
Non-African Service 2003 2009 growth rate (%)
Purpose of visit visitors African visitors Hotel rooms 650 4,256 37
Leisure 1,623 1,136 Restaurants 50a 94 17
Conference and Tour operator companies 12 26 14
business 1,623 108 Travel agencies 5 24 30
Visiting friends and
relatives 120 84 Source: RDB/ORTPN.
Transit/other 119 83 a. Data from 2005.

Source: RDB/ORTPN/MINECOFIN, based on visitor expenditure

survey from 2006, adjusted for inflation.
domestic tourists was two days, while the most frequently
cited length of stay by international tourists was four days.
The number of hotel rooms and tour operators in
Rwanda has increased significantly in recent years, under-
scoring the successful recovery of the tourism sector. Data
available from RDB and ORPTN show that the number of
hotel rooms increased from 650 in 2003 to 4,256 in 2009,
more than 500 percent overall and 37 percent annually on
average (table 14.3). Information on occupancy rates, how-
ever, is very limited. The only available information spans
January to March 2008 and indicates an average room occu-
pancy rate of 36 percent. There was a large difference in the
occupancy rates by the grade of accommodation. Room
occupancy rates for upper-grade accommodation (a total of
453 rooms) were more than 70 percent on average, whereas
rates for lower-grade accommodation (2,264 rooms) were
only 28.5 percent (ORTPN 2008c, 2008d).
A total of 26 tour operators were active in Rwanda as of
2009, compared to none in the 1980s. The number of
restaurants and travel agencies has grown as well. Most
Rwandan tour operators started out with little available
finance, which limited their possibilities. They could, for
example, not afford to buy a car, but had instead to rent cars
as they needed them. Most of those tour operators have suc-
cessfully expanded their businesses by now, though, and
some regional operators have also opened offices in Kigali.
The tourism sector has also experienced a significant
amount of privatization. The situation is much different
from the 1980s, when all hotels except one were government
owned. Although the government still held a share in two
hotels in 2010, it was not involved in their management,
leaving the tourism sector almost entirely in the hands of

244 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
the private sector. The government still owns and runs the
national parks, although in 2009 a concession was awarded
to the private company African Parks to manage Akagera
National Park.
Tourism income, as recorded by ORPTN, has increased
in recent years. The majority of revenue is derived from
VNP entrance and gorilla permit fees. Other income is
raised through the gorilla naming ceremony, partners and
donors, and interest from treasury bills and other income.
Most of the revenue is spent on operating expenditures,
whereas capital expenditures paid out of the budget of
ORPTN are limited. Large investments are funded by the
overall central government budget. Revenue shared with
communities has increased since 2005, representing 8 per-
cent of total operating expenditure in 2008, up from 6 per-
cent in 2006.
Foreign investment in Rwanda’s tourism sector is sub-
stantial. Between 2000 and 2009, foreign direct investment
of RF 258 billion went into hotels and leisure, accounting
for 20 percent of total FDI inflows into the country (figure
14.8). Local investment in hotels, restaurants, and other
tourism activities amounted to RF 140 billion between 1999
and 2009, representing 16 percent of total local investment
over that period.
Although the amount of foreign investment exceeded the
amount of local investment, it has been concentrated on a
small number of projects. Local investors play an important
Figure 14.8 FDI to Rwanda, by Sector, 2000–2009
Hotels and
leisure Mining and
20% energy
33%
Other
24%
Telecom
15%
Construction and
real estate
24%
Source: RDB; included are RDB registered investments.
CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
role in the development of the tourism sector in Rwanda.
Eighty-six percent of all new projects since 1999 that are
operational were financed by local investors. Moreover, a
group of private investors has established the Rwanda
Investment Group (RIG) to pool resources. Several sub-
RIGs have also been created, one of which is currently plan-
ning the construction of a convention center intended to
hold up to 2,000 participants, which would increase
Rwanda’s chance of attracting large conferences.4
Finally, the structure and organization of the tourism sec-
tor has been reformed to assign clear responsibilities. The
Ministry of Commerce, Industry, Investment Promotion,
Tourism, and Cooperatives (MINICOM) holds overall
responsibility for tourism. The private sector is represented
by the tourism chamber, which consists of four industry
associations: accommodation, tour operators, transport, and
private education establishments. The tourism chamber is,
however, still supported by the government because of insuf-
ficient resources. It is part of the private sector federation,
the equivalent of a chamber of commerce and industry.
Contribution of tourism to the economy
Rwanda has made remarkable progress in terms of eco-
nomic growth since the genocide in 1994 (figure 14.9).
Growth averaged 15.6 percent in the five years after the
genocide in 1994, declined to an average of 6.6 percent
between 2000 and 2004, and increased again to an average
of 8.4 percent between 2005 and 2008. A significant increase
in GDP per capita has been recorded along the way, from
$142 in 1994 to $313 in 2008 (both in 2000 prices).
While the contribution of tourism to GDP remains
small, the sector has become Rwanda’s main source of
export revenue. The category “restaurants and hotels” has
contributed less than 2 percent to overall GDP and 4 per-
cent to the services sector on average since 1999, but value
added from restaurants and hotels (at constant prices) has
recorded a steady increase of 21 percent on average.5 The
main increase in the services sector came from wholesale
and retail trade and other services (education, health,
finance and insurance, and real estate). The measurement of
tourism’s contribution to GDP, however, is difficult, because
transport services, for example, constitute a large share of
tourism revenues but are not included in the “hotel and
restaurants” category.
Overall, exports of nonfactor services from Rwanda have
outperformed exports of goods (such as coffee and tea) as
Rwanda’s main foreign exchange earner (figure 14.10).
Travel is the largest component of exports of nonfactor
245
 Figure 14.9 Growth in Real GDP and GDP per Capita, 1993–2008

40

30 300

20

Constant 2000 USD

10 250

Percent
0

–10 200

–20

–30 150

–40

–50 100
93
94
95
96
97
98
99
00
01
02
03
04
05
06
07
08
19
19
19
19
19
19
19
20
20
20
20
20
20
20
20
20
Real GDP (left axis) Per capita GDP (right axis)

Source: International Monetary Fund.

Figure 14.10 Comparison of Composition of Export of Goods and Nonfactor Services

(Averages)

a. 1995–98 b. 2005–08
Nonfactor
Nonfactor
services excl.
services excl. Coffee
travel
travel 11% Tea
20%
Coffee 23% 8%
37%

Travel Other
20% goods
Travel
Other Tea 22%
36%
goods 11%
12%

Source: IMF.

services; other categories include other transportation and
freight and insurance. After a decline from $19 million in
1993 to only $6 million in 1995, revenue from tourism
increased tremendously to $202 million in 2008 (figure
14.11). A comparison with the preconflict period is diffi-
cult, because data on tourism revenue are available only for
1992 onward.
Rwanda has identified tourism in its Economic Devel-
opment and Poverty Reduction Strategy as a national pri-
ority sector to eradicate poverty (Republic of Rwanda
2007). According to rough estimates, each of the three big
business hotels in Kigali6 generates about $500,000 per year
in income for semiskilled and unskilled workers, food pro-
ducers, and artisans. Tourists visiting the VNP and the
Musanze area generate around $1 million in income for
poor workers and producers. In addition, the area receives
large amounts of donations and grants7 (SNV and ODI
2008). As much as possible, hotels source their supplies
(particularly food products) from the local market to con-
tribute to the economy.
Estimates for 2009 indicate that the tourism industry
directly employs 33,800 people in Rwanda, whereas indirect
employment accounts for another 40,500 jobs, resulting in
total travel and tourism–related employment of 74,300 jobs.
Tourism thus represents 4.0 percent of total employment
in Rwanda, only slightly below the Sub-Saharan African

246 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
Figure 14.11 Tourism Revenue, 1992–2008

220
200
180
160

U.S.$ (millions)
140
120
100
80
60
40
20
0
92

93

94

95

96

97

98

99

00

01

02

03

04

05

06

07

08
19

19

19

19

19

19

19

19

20

20

20

20

20

20

20

20

20
Tourism revenue (current prices) Tourism revenue (constant 2005 prices)

Source: IMF and World Bank, World Development Indicators database.

average of 4.6 percent, and well below the amount in Kenya
(7.1 percent), Tanzania (7.1 percent), and Uganda (6.6 per-
cent). Employment in the tourism industry (direct and
indirect) has grown by 2.8 percent on average in Rwanda
over the past 10 years, compared with 3.4 percent in Kenya,
2.5 percent in Tanzania, and 2.4 percent in Uganda (World
Travel and Tourism Council 2009).
REMAINING CONSTRAINTS AND EMERGING
POSSIBILITIES FOR TOURISM
Despite the good performance of the tourism sector in
Rwanda, several challenges remain. The main impediment
cited by almost all actors in the sector is the large skill deficit.
This deficit applies to all areas of tourism, including guides,
chefs, and hotel service personnel and technicians. Hotels
and tour operators either train their staff in house or send
them to neighboring countries to be trained, although
recently some tourism schools have been opened by the pri-
vate sector. The emphasis of the curriculum is on managerial
rather than technical skills, meaning that the demands of the
sector are not taken into account adequately. To accomplish
the goal of turning Rwanda into a service-oriented economy,
skill development is of utmost importance.
Several other challenges confront the tourism sector:
■ There is an overreliance on gorilla tourism. The number
of permits cannot be easily increased, and the existing
permits are typically sold out. Tourism needs to be
diversified and other attractions promoted, such as bird-
ing and primates in Nyungwe, visits to Lake Kivu, and
conference tourism. Diversity is particularly important
in establishing Rwanda as a stand-alone destination.
■ Access to finance is still an impediment for the devel-
opment of the sector. Banks seem reluctant to finance
tourism projects because they are within a service-
oriented sector rather than attached to a sector produc-
ing tangible goods. Stakeholders in the tourism sector
have proposed the establishment of a guarantee fund by
the government.
■ Other sectors need to be further promoted through the
tourism sector in order to reduce poverty. Pro-poor links
that can be exploited further include the food supply
chain to hotels, lodges, and restaurants; assistance to
poor households to access training, employment, and
promotion in hospitality; practical initiatives to help
businesses enhance their own business models; and part-
nerships with more domestic and regional tour opera-
tors, hotels, and lodges to promote community activities
such as cultural events and the sale of handicrafts (SNV
and ODI 2008).
■ Rwanda needs to comply with international standards.
Currently, hotels are being classified according to the EAC
standard. While this is an important step, it poses chal-
lenges to a number of hotels. Once Rwanda shifts to an
international system of standards and classifies hotels
accordingly, the pricing structure might have to be revised
to remain competitive with the neighboring countries.

CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 247
 ■ RDB and MINICOM do not have sufficient staff. In
MINICOM, only one part-time person is responsible for
tourism. RDB, as well, has only a limited number of staff.
Given the very ambitious agenda for the tourism agenda,
sufficient staffing should be in place.
■ Infrastructure needs improvement. Although roads in
Rwanda are broadly adequate, the air transport con-
nection to the country is still limited. A new airport is
currently being planned, which could attract more
international carriers to offer direct flights, especially
from Europe.
In the face of the challenges, several possibilities are
emerging to diversify the tourism sector and increase its
contribution to the economy. Regarding leisure tourists,
primate tours, and birding in the Nyungwe Forest are the
most promising areas of diversification. New experiences
would extend the length of stay of tourists and eventually
establish Rwanda as a stand-alone destination. Lessons
learned from gorilla tourism regarding conservation, how-
ever, should be taken into account when developing tourism
attractions in other national parks. The development con-
cept of Rwanda’s Destination Management Areas is identi-
fied in the Sustainable Tourism Development Master Plan
(Republic of Rwanda 2009b). Conference tourism could
also be of great benefit to Rwanda. In this regard, the con-
struction of a conference center that will accommodate up
to 2,000 people is under way. Further, a draft action plan for
the development of MICE (meetings, incentives, confer-
ences, and exhibitions) tourism has been prepared by the
Tourism Working Group.
NOTES
1. EAC member countries are Burundi, Kenya, Rwanda,
Tanzania, and Uganda.
2. The gorillas are visited by independent travelers, over-
landers, and high-end tours (ORTPN 2004).
3. The optimum group size, according to tourists, is six
people; tourists are not willing to pay more to reduce group
size beyond this (Fawcett 2009; Weber 1993).
4. The currently available conference facilities can host a
maximum of 500 and 1,000 participants.
5. The share of hotels and restaurants is not available for
years before 1999.
6. Serena Hotel Kigali, Hotel de Mille Collines, and Laico
Umubano Kigali Hotel (former Novotel Kigali).
7. Including contributions to community development
from ORTPN’s revenue-sharing program and donations by
tourists.
248 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
BIBLIOGRAPHY
Barrett, C., and P. Arcese. 1995. “Are Integrated Conservation-
Development Projects Sustainable?” World Development
23 (supplement 7): 1073–84.
Bush, G. 2009. Personal Communication, Woods Hole
Research Center, November 30.
Bush, G., and K. Fawcett. 2007. “An Economic Study of
Mountain Gorilla Tourism in the Virunga Volcanoes
Conservation Area.” U.S. Fish and Wildlife Service, Dian
Fossey Gorilla Fund International, Ruhengeri, Rwanda.
Bush, G., N. Hanley, and S. Colombo. 2008. “Measuring the
Demand for Nature-based Tourism in Africa: A Choice
Experiment Using the ‘Cut-off ’ Approach.” Discussion
Paper 2008-06, Stirling Economics, University of Stirling,
Stirling, U.K.
Butynaski, T. M., and J. Kalina. 1997. “Gorilla Tourism: A
Critical Look.” Africa Biodiversity Conservation Pro-
gram, Nairobi.
Fawcett, K. 2009. Personal Communication, Director,
Karisoke Research Centre, September 16.
Fawcett, K., C. Hodgkinson, and P. Mehlman. 2004. “An
Assessment of the Impact of Tourism on the Virunga
Mountain Gorillas: Phase I—Analyzing the Behavioral
Data from Gorilla Groups Designated for Tourism.” Dian
Fossey Gorilla Fund International, Atlanta, GA.
Gray, M., et al. 2005. “Virunga Volcanoes Range and Census,
2003.” Joint organizers report, Uganda Wildlife Author-
ity/ORTPN/Institut Congolaise pour la Conservation du
Nature.
Grosspietsch, M. 2006. “Perceived and Projected Images of
Rwanda: Visitor and International Tour Operator Per-
spectives.” Tourism Management 27: 225–34.
Homesy, J. 1999. “Ape Tourism and Human Diseases; How
Close Should We Get?” Technical Report, International
Gorilla Conservation Program, Nairobi.
Hulme, D., and M. Murphee. 2001. “Community Conser-
vation in Africa.” In African Wildlife and Livelihoods: The
Promise and Performance of Community Conservation,
ed. D. Hulme and M. Murphee. Oxford, U.K.: James
Curry Ltd.
IFC (International Finance Corporation). 2007. “Rwanda
Hotel Market Study.” Washington, DC.
Kalpers, J., et al. 2003. “Gorillas in the Crossfire: Population
Dynamics of the Virunga Mountain Gorillas over the
Past Three Decades.” Oryx 37 (3): 326–37.
Litchfield, C. A. 1997. “Treading Lightly: Responsible
Tourism with the African Great Apes,” Travellers’ Medical
and Vaccination Centre, Adelaide. Cited in Fawcett,
Hodgkinson, and Mehlman (2004).
Makambo, W. 2009. “Sabyinyo Silverback Lodge: A Com-
munity Partnership for Conservation.” Presentation
 from RDB-TC tourism forum, International Gorilla
Conservation Program, Nairobi.
Mehta, H., and C. Katee. 2005. Viriunga Massif Sustainable
Tourism Development Plan, DR Congo, Rwanda,
Uganda. ECDA: Fort Lauderdale FL.
Murnyak, D. F. 1981. “Censusing the Gorillas in Kahuzi-
Biega National Park.” Biological Conservation 21: 163–76.
OTF Group. 2008a. “Building a Winning MICE Tourism
Product—Target Customer Needs.” Rwanda Tourism
Workgroup.
———. 2008b. “Building a Winning Bird Product—Target
Customers and Channel Partner Needs.” Rwanda
Tourism Workgroup.
ORTPN (Office Rwandaise du Tourisme et des Parcs
Nationaux). 2004. “Strategic Plan 2004–2008.” Draft ver-
sion 2, Kigali.
———. 2008a. “Highlights on National Parks Visitation in
Rwanda—2008.” Vol. l.1, Kigali.
———. 2008b. “Highlights of International Visitor Arrivals
in Rwanda—2008.” Kigali.
———. 2008c. “Highlights on Accommodation Statistics in
Rwanda—January 2008.” Vol. I., Kigali.
———. 2008d. “Highlights on Accommodation Statistics in
Rwanda—March 2008.” Vol. I.2, Kigali.
Plumptre, A. et al. 2004. “The Socio-economic Status of
People Living Near Protected Areas in the Central Alber-
tine Rift.” Albertine Rift Technical Reports 4, Albertine
Rift, Program, Wildlife Conservation Society, New York.
Republic of Rwanda. 2002. “A Plan to Unleash Rwanda’s
Tourism Potential.” Kigali.
———. 2007a. “Economic Development and Poverty
Reduction Strategy, 2008s–2012.” Ministry of Finance
and Economic Planning, Kigali.
———. 2007b. “A Revised Tourism Strategy for Rwanda—
Sustaining the Momentum.” Ministry of Commerce,
Industries, Investment Promotion, Tourism and Cooper-
atives, Kigali.
———. 2009a. “Rwanda Tourism Policy.” Ministry of Trade
and Industry, Kigali.
———. 2009b. “Sustainable Tourism Development Master
Plan for Rwanda—Final Report.” Project of the Republic
of Rwanda and the United Nations World Tourism
Organization.
Rutagarama, E., and A. Martin. 2006. “Partnerships for Pro-
tected Area Conservation in Rwanda.” Geographical Jour-
nal 172 (4): 291–305.
Sabuhoro, E. 2009. “Contribution of Tourism to Poverty
Reduction in Rwanda: A Tour Operator’s Perspective.”
Presentation at the Delphe/EPA Workshop, July 6–7.
CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE
SNV Rwanda and ODI. 2008. “Creating Pro-poor Linkages
around Rwanda Tourism.” Kigali.
SNV Rwanda and RDB. 2009. “Tourism Market Research
for Rwanda 2009: A Survey of Tourists, Tour Operators
and Accommodation—Final Report.” Kigali.
Spenceley, A., S. Habyalimana, R. Tusabe, and D. Mariza.
2010. “Benefits to the Poor from Gorilla Tourism
in Rwanda.” Development Southern Africa, 27 (5):
647–62.
Télesphore, Ngoga. 2009. Interview. September. RDB,
Kigali.
Townsend, C. Forthcoming. “Guidelines for Community-
based Tourism in Rwanda.” ORTPN / UNWTO ST-EP.
(Sustainable Tourism–Eliminating Poverty).
Tusabe, R., and S. Habyalimana. 2010. “From Poachers to
Park Wardens: Revenue Sharing Scheme as an Incentive
for Environment Protection in Rwanda.” Mountain
Forum Bulletin (10) 1.
Uwingeli, P. 2009. Personal Communication, Chief Park
Warden, Volcanoes National Park, November 3.
Vedder, A., and B. Weber. 1990. “The Mountain Gorilla
Project.” In Living With Wildlife: Wildlife Resource Man-
agement with Local Participation in Africa, ed. A. Kiss.
Washington, DC: World Bank.
Verdugo, D. 2009. “Challenges for Developing Community
Understanding of Tourism in an Emerging Destina-
tion.” Unpublished paper cited in Spenceley, A. et al.
(2010).
Weber, B. 1982. “People, Gorillas...or Both?” London: Earth-
scan.
———. 1985. “Le Parc National des Volcans Biosphere
Reserve, Rwanda: The Role of Development in Conser-
vation.” Parks 10 (3): 19–21.
———. 1987. “Socioecologic Factors in the Conservation of
Afromontane Forest Reserves.” In Primate Conservation
in the Tropical Rainforest, ed. C. Marsh and R. Mitter-
meier, 205–29. New York: Alan R. Liss, Inc.
———. 1993. “Ecotourism and Primate Conservation in
African Rain Forests.” In The Conservation of Genetic
Resources, ed. C. Potter, 129–50. Washington, DC: AAAS
Press.
Woodford, M., T. Mutynsky, and W. Karesh. 2002. “Habitu-
ating the Great Apes: The Disease Risks.” Oryx 36 (2):
153–60.
World Bank. 2008. “Country Assistance Strategy for the
Republic of Rwanda for the Period FY09-FY12.” Report
No. 44938-RW, Washington, DC.
World Travel and Tourism Council.s 2009. “Travel and
Tourism Economic Impact—Rwanda 2009.” London.
249
PA R T I V

Boosting Agricultural Efficiency and Output

through Targeted Interventions
 CHAPTER 15
Increasing Rice Productivity and
Strengthening Food Security through New
Rice for Africa (NERICA)
Aliou Diagne, Soul-Kifouly Gnonna Midingoyi, Marco Wopereis, and
Inoussa Akintayo
espite Africa’s potentially rich land and water
D
resources, its farmers are among the poorest in the
world. Because the vast majority of people in
Africa derive their livelihoods from agriculture, the weak
state of the sector has profound implications for poverty.1
Agricultural innovation in Africa needs to internalize the
region’s biophysical, institutional, and socioeconomic con-
straints and establish efficient value chains to support sus-
tainable growth and reduce poverty.
Agricultural research can catalyze agricultural innova-
tion and the development of the value chain. A prime exam-
ple is the New Rice for Africa (NERICA) varieties developed
by the Africa Rice Center (AfricaRice2) and partners, which
gained popularity among Africa’s rice farmers in a relatively
short period of time.3 NERICA varieties were developed
from crosses between the African species Oryza glaberrima
Steud. and the Asian species O. sativa L., using conventional
biotechnology to overcome the sterility barrier between the
two species. This chapter describes the history, develop-
ment, and impact of NERICA and makes recommendations
about its expansion.
THE IMPORTANCE OF RICE AND THE
DEVELOPMENT OF NERICA VARIETIES
Rice has long been the food staple in many traditional rice-
growing communities and in major cities in West Africa.4 It
is now the fastest-growing food staple in Africa. Annual per
capita rice consumption in West Africa increased from
14 kilograms in the 1970s to 22 kilograms in the 1980s and
more than 39 kilograms in 2009. For Africa as whole, annual
per capita rice consumption increased from 11 kilograms in
the 1970s to 21 kilograms in 2009 (figure 15.1). Since the
early 1970s rice has been the number one source of caloric
intake in West Africa and the third most important source
of calories (after maize and cassava) for the continent as a
whole (figure 15.2).
Domestic rice production grew at the rate of 6 percent a
year between 2001 and 2005.5 But production still falls far
short of demand. As a result, Africa imports up to 40 per-
cent of its rice consumption. The continent as a whole
imports 30 percent of world rice imports, a potentially very
risky and unsustainable situation, as shown during the food
crisis in 2008. Rice imports cost Africa almost $4 billion in
2009 (Seck et al. 2010)—money that could have been
invested in developing the domestic rice sector.
Until the early 1990s, rice breeding programs in Africa
worked almost exclusively on improving O. sativa lines
for Africa’s diverse rice growth environments. Some
breeders even voted to accelerate the disappearance of
glaberrima (IRAT 1967), even though sativa varieties are
generally much more vulnerable than glaberrima to the
numerous biotic stresses (rice diseases, insect attacks,
weeds, and so forth) and abiotic stresses (soil acidity,
drought, salinity, iron toxicity, and so forth) of the
African environment.
253
 Figure 15.1 Annual per Capita Rice Consumption in Africa and West Africa, 1961–2009

45

40

35
West Africa
Kilograms/person/year 30

25

20

15
Africa
10

0
19 1
63
65
67
69
71
73

19 5
77
79

19 1
83
85
87
89

19 1
93

19 5
97
99
01
03

20 5
07
09
6

0
19

19
19
19
19
19
19

19
19

19
19
19
19

19

19
20
20
20

20
Source: FAO 2009.

Figure 15.2 Share of Calories in African Diet Provided by Various Sources, 1961–2007

18

16

14
Percent of total calorie intake

12

10

0
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
95
97
99
01
03
05
07
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
20
20
20
20

Cassava Maize Millet Rice (milled equivalent) Sorghum

Source: FAO 2009.

Some visionary breeders thought about combining the ing breeders from the Institut de Recherche Agronomique
favorable characteristics of O. sativa with those of O. glaber- Tropicale (IRAT) and the Institute of Savannah (IDESSA)
rima. Crossing was difficult, however, because of the steril- Bouaké, in Côte d’Ivoire, worked on the problem during the
ity barrier between the two species. A few breeders, includ- 1970s and 1980s (Diagne 2006). Their work did not produce

254 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
good results, with the crosses resulting in either sterile off-
spring or offspring that had some of the undesirable char-
acteristics of the glaberrima parent (lodging, shattering, low
yield).6
Development of the NERICA varieties began in 1991,
when AfricaRice initiated an interspecific breeding program
for the upland ecosystem. This long-term investment paid
off, when breeders eventually overcame the obstacles
encountered earlier, through perseverance and the use of
biotechnology tools such as anther culture and embryo res-
cue techniques. In 1994 the first interspecific line with
promising agronomic performance was obtained (Wopereis
et al. 2008; Jones et al. 1997a, b).7 Several hundred interspe-
cific progenies were generated, opening new gene pools and
increasing the biodiversity of rice.
The main objective of the breeding work that led to the
NERICAs was to combine the high-yielding attribute of
O. sativa with the resistance of the indigenous O. glaberrima
to the African environment. Another long-sought attribute
for a good upland variety is the ability to provide acceptable
yields under the low-input use conditions typical of upland
rice farming in Africa. Both objectives have been largely met,
as evidenced by the experimental trials data summarized in
Box 15.1 The Participatory Varietal Selection Methodology
Participatory varietal selection (PVS) aims to improve
rice production through direct involvement of farmers,
who play an active role in varietal selection, develop-
ment, and dissemination. PVS is an adjunct to conven-
tional breeding that markedly reduces the time taken
for new varieties to reach farmers. In conventional
breeding schemes, selection and testing procedures
involve a series of multilocation trials over 8 to 12 years
using a diminishing number of varieties. Through PVS,
a large number of potentially interesting varieties reach
farmers for evaluation in 5 years, and farmers have a
major input into the selection of varieties to be released.
By involving farmers in the varietal selection process,
PVS aims to meet a diversity of demands for varieties,
taking account of site-specificity for agronomic adapta-
tion and selection criteria.
AfricaRice’s approach to PVS research involves a
three-year program. In the first year, breeders identify
centralized fields near villages and plant a rice garden
trial of up to 60 upland varieties. The varieties range
from traditional and popular O. sativas to NERICAs,
Source: Gridley and Sié 2008.
CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
table 15.A1 and figure 15.A1 in annex 15.A1, which compare
the performance of NERICA progenies with that of their
sativa and glaberrima parents and other sativa checks under
high and low input conditions and under major stresses in
upland rice ecologies in AfricaRice’s key research sites in
Cote d’Ivoire.8
Indeed, the agronomic trials data show that the NERICA
varieties bring to upland farmers the high yield potential
of the sativa varieties under both low and high input con-
ditions (see figure 15.A1), with what is in essence insur-
ance against the risk of significant yield losses in the face
of major upland biotic and abiotic stresses. Given that the
sativa and glaberrima parents of these first-generation
NERICAs are not the best-performing varieties within the
two species, there is reason to expect that the perfor-
mance of the next generations of NERICAs will be much
higher.9
Through participatory varietal selection (box 15.1),
farmers all over Africa have evaluated interspecific lines.
The most successful lines have been named NERICA vari-
eties. There are currently 18 varieties suited for upland
growth conditions (NERICA1 to NERICA18) and 60 vari-
eties suited for lowland growth conditions (NERICA-L1 to
African O. glaberrimas, and local varieties as checks.
Male and female farmers are invited to visit the plot
informally as often as possible. Groups of farmers for-
mally evaluate the varieties at three key stages: maxi-
mum tillering (development of the plant leaves), matu-
rity, and postharvest. At the first two stages, farmers
compare agronomic traits. Each farmer’s varietal selec-
tion and the criteria for selection are recorded and later
analyzed.
In the second year, each farmer receives as many as
six of the varieties he or she selected in the first year to
grow. PVS observers, including breeders and techni-
cians from nongovernmental organizations and exten-
sion services, visit participating farmers’ fields to record
seed performance and farmer evaluations of the
selected varieties. At the end of the year, farmers evalu-
ate threshability and palatability to provide an overall
view of the strengths and weaknesses of the selected
varieties. In the third year, farmers are asked to pay for
seeds of the varieties they select. Willingness to pay pro-
vides evidence of the value farmers place on the seeds.
255
 NERICA-L60). Agronomic characteristics of these vari-
eties vary widely, but they are generally high yielding and
early maturing (a trait much appreciated by farmers), do
not lodge or shatter, yield good grain quality, and are
relatively resistant to the biotic and abiotic stresses of the
continents harsh growth environment. Weed resistance of
the NERICAs, especially the upland NERICAs, still needs
improvement. (For details on the development and the char-
acteristics of different NERICA varieties, see Somado, Guei,
and Keya 2008; Wopereis and others 2008; and Kaneda
2007a, b, c, and d.)
ADOPTION AND IMPACT OF NERICA VARIETIES
IN SUB-SAHARAN AFRICA
The first generation of interspecific progenies was intro-
duced through participatory varietal selection trials in Côte
d’Ivoire in 1996 and in other member countries of
AfricaRice in 1997. NERICA1 and NERICA2 were officially
released in 2001 in Côte d’Ivoire (Diagne 2006; Science
Council 2007). NERICA varieties have now been introduced
in more than 30 countries in Sub-Saharan Africa using PVS
approaches, mostly thanks to the African Rice Initiative
(ARI).10 By 2007, 17 upland NERICA varieties and 11 low-
land NERICA varieties had been adopted. Where needed,
ARI produced foundation seed to stimulate uptake of NER-
ICA and other improved varieties. Seed production by ARI
has increased steadily, from 2,733 tons in 2005 to 7,238 tons
in 2007 and 13,108 tons in 2008 (Akintayo et al. 2009).
About 700,000 hectares in Sub-Saharan Africa were esti-
mated to be under NERICA varieties as of 2009, covering an
estimated 5 percent of the continent’s upland rice-growing
areas. This level of penetration is impressive 8–10 years after
release of the first varieties. The studies reviewed in this sec-
tion describe the areas under NERICA varieties in selected
countries, actual and potential rates of adoption among
farmers, and the impact of NERICA varieties on rice pro-
ductivity and the livelihoods of family farmers.
Estimated area under NERICA
Surveys conducted by national agricultural research systems
provide estimates of the total area under NERICA varieties
in different countries. Diagne et al. (2006) and Adegbola et
al. (2006) estimate the total area under NERICA varieties at
51,000 hectares in Guinea in 2004 and 5,000 hectares in
Benin in 2003. Based on its estimates of the quantities of
seed produced and distributed to farmers, the National
Food Reserve Agency (NFRA) of Nigeria reports 186,000
256 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
hectares under NERICA1 in Nigeria in 2007 (ARI 2008).
The Uganda National Research Institute, citing a report of
the Statistics Office of the Ministry of Agriculture in
Uganda, estimates that 35,000 hectares were under NER-
ICA4 in 2007 (ARI 2008).
Except in Guinea and Benin, where surveys were con-
ducted (in 2003 in Guinea and in 2005 in Benin), the esti-
mated areas under NERICA reported above by national
agricultural research systems are based on quantities of seed
produced and distributed to farmers. To meet the need to
quantify uptake of NERICA and other improved varieties in
a much more reliable manner, AfricaRice and partners are
currently conducting surveys in many countries where
NERICA is cultivated. Surveys conducted in 2009 by
national agricultural research systems and national agricul-
tural statistical services in 21 Sub-Saharan African countries
provide more recent estimates for some countries
(AfricaRice 2010). According to these data, the area under
NERICA was about142,391 hectares in Guinea and 244,293
hectares in Nigeria in 2009.11
Adoption of NERICA varieties
A 2009 study on adoption in Sub-Saharan African countries
estimates actual rates of adoption of NERICA varieties
(table 15.1).12 These rates illustrate a largely unrealized
potential adoption rate if the full rice farming population of
these countries could be exposed to NERICA varieties and
provided access to seed.
Uptake of NERICA varieties was reportedly high in the
Kaduna and Ekiti states of Nigeria (Spencer et al. 2006).
About 30 percent of farmers in Ekiti cultivated NERICA1 in
2005; 42 percent of farmers in PVS villages and 19 percent
in near–PVS villages did so in Kaduna.13 Adoption of NER-
ICA1 appears to have continued during 2004 and 2005,
despite the scaling down of PVS activities. About 35 percent
Table 15.1 Estimated Actual and Potential Rates of
Adoption of NERICA Varieties in
Selected African Countries
(percent)
Rate of Potential rate
Country (year) adoption of adoption
Benin (2004) 19 47
Côte d’Ivoire (2000) 4 24
Guinea (2001) 20 61
The Gambia (2006) 40 87
Source: Diagne 2009.
of farmers in near–PVS villages in Ekiti and Kaduna had not
Table 15.2 Yields and per Capita Income in
heard about NERICA1, showing the potential for increase in Selected African Countries, by Gender
adoption rates (Spencer et al. 2006). In 2005 an estimated
14 percent of farmers in PVS villages in Kaduna and 9 percent Yield
of farmers in near–PVS villages had adopted NERICA2. Country (year) (tons/hectare) Income ($)

The percentage of households that grew NERICA vari- Benin (2004)

eties in Uganda increased from 0.9 percent in 2002 to 2.9 Men 0.26*** 22***
Women 0.97*** 38***
percent in 2003 and 16.5 percent in 2004, according to All 0.71*** 25***
Kijima, Otsuka, and Sserunkuuma (2008). They find that Côte d’Ivoire (2000)
membership in a farmers group, formal education of the Men 0.44 —
Women –0.45 —
head of household, and the number of household members All ^0.44 —
significantly increased the probability of adopting NERICA The Gambia (2006)
varieties as well as the scale of area planted to NERICA vari- Men 0.18 4.6
Women 0.142*** 10.6**
eties. Land size per person had a negative effect on the share All 0.146*** 10.2**
of land planted to NERICA varieties, suggesting that land-
Source: Authors compilation, based on data from Agboh-
poor households tend to allocate a larger proportion of land Noameshie, Kinkingninhoun-Medagbe, and Diagne (2007) for
to cultivation of NERICA varieties. This pattern is also Benin, Diagne; Midingoyi, and Kinkingninhoun (2009) for Côte
observed among households headed by women, who are d’Ivoire; Dibba et al. (2008a, b) for The Gambia; Dontsop et al.
(2010) for Nigeria; and Kijima, Otsuka, and Sserunkuuma (2008)
likely to be poorer. Poorer households in Uganda thus tend for Uganda.
to allocate a larger proportion of their land to NERICA vari- — = Not available.
eties, which may suggest that the adoption of NERICA has *** Significant at the 1 percent level; ** significant at the
5 percent level.
the potential to reduce poverty and improve income distri-
bution as it increases rice yields.

Impact of NERICA varieties on productivity and
livelihoods of farmers
NERICA rice varieties were developed with the aims of
improving rice productivity, raising income, and reducing
food insecurity of poor upland rice farmers (mostly
women), who rarely use fertilizer, which they say they can-
not afford. Empirical evidence from impact assessment
studies conducted in West Africa points to a heterogeneous
impact of NERICA adoption across and within countries
(table 15.2). Impacts have been significantly positive in
Benin and Gambia. In almost all countries, the impacts have
generally been higher for women than for men (Diagne,
Midingoyi, and Kinkingninhoun 2009).
In Benin farmers adopting NERICA enjoyed an addi-
tional yield gain of about 1 metric ton per hectare. The
impact at the national level was very limited, however,
because of the low diffusion of NERICA varieties in Benin
(Adegbola et al. 2006).
Results from another analysis based on data from the
2004 season show that the impact of NERICA adoption is
higher for women than for men (Agboh-Noameshie, Kink-
ingninhoun-Medagbe, and Diagne 2007). Female potential
adopters have a surplus of production of 850 kilograms of
paddy per hectare, compared with 517 kilograms of paddy
per hectare for men; the potential net income gain from
adopting NERICA is estimated at about $337 per hectare for
women and $277 for men. Results in The Gambia indicate
that an additional yield gain of 0.14 ton per hectare was
achieved by rice farmers, most of them women, adopting
NERICA varieties (Dibba 2010).14 Results from Côte
d’Ivoire show that the impact is heterogeneous, with a siz-
able and statistically significant impact found for female
farmers (0.7 ton per hectare) and a nonstatistically signifi-
cant impact found for male farmers (Diagne 2006; Diagne,
Midingoyi, and Kinkingninhoun 2009).
In Uganda NERICA had positive effects on productivity
and allowed farmers to improve their yields (Kijima, Sserunk-
uuma, and Otsuka 2006; Kijima, Otsuka, and. Sserunkuuma
2008; Bergman-Lodin 2005). According to Kijima, Sserunku-
uma, and Otsuka (2006), the average yield of NERICA in
Uganda was 2.2 tons per hectare—twice the average rice yield
in Sub-Saharan Africa. They find large differences in yields
between experienced (2.46 tons per hectare) and inexperi-
enced (1.72 tons per hectare) households, indicating that
experience matters in achieving high yields. Other yield
determinants were rainfall and cropping patterns.
A few studies document the impact of adoption of NER-
ICA varieties on household income and poverty. Most find
that adoption increases household income and improves

CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 257
 income distribution within households. In Uganda, for
example, adoption of NERICA varieties has the potential to
increase annual per capita income by $20 (12 percent of
actual per capita income) and to reduce the incidence of
poverty, measured by the headcount ratio, by 5 percentage
points (from 54.3 percent to 49.1 percent) (Kijima, Otsuka,
and Sserunkuuma 2008). The poverty gap index and the
squared poverty gap index also declined following the intro-
duction of NERICA, suggesting that adoption can increase
incomes among the absolute poor in Uganda.
NERICA adoption has had a positive impact on house-
hold rice and total incomes in The Gambia (Glove 2009;
Dibba 2010). Dibba (2010) finds that adoption of NERICA
increases rice farmers’ daily income by about $0.34 on aver-
age (about 30 percent of the average daily income of
adopters). Sogbossi (2008) identifies NERICA adoption as a
key determinant in poverty reduction in Benin, where it
reduced the probability of being poor by 13 percent.15 The
impact of NERICA adoption was higher for women farmers
(reduction in probability of being poor was 19 percent)
than for men (reduction in probability of being poor was
6 percent). Findings from a study conducted in Nigeria indi-
cate that adoption of NERICA increased total farm house-
hold gross income from rice by about $555 (about 46 percent
Box 15.2 Successfully Disseminating NERICA Varieties in Burkina Faso, The Gambia, Guinea, and Uganda
Partnerships with the National Agricultural Research
Extension Systems, nongovernmental organizations,
and farmer organizations, with support from donors
and development partners, were key ingredients in scal-
ing up the dissemination of NERICA varieties through-
out Sub-Saharan Africa.a Governments across the
region also supported the dissemination of NERICA
varieties through major rice development initiatives
well before the 2008 food crisis.
Burkina Faso
Domestic rice production in Burkina Faso increased by
an astonishing 241 percent in 2008 (Solidarité et Progres
2008), partly as a result of the adoption of NERICA vari-
eties. This remarkable achievement reflected the govern-
ment’s response to the world food crisis in 2008. Emer-
gency support facilitated farmers’ access to mineral
258 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
of average gross income from rice), increasing farmer’s prob-
ability of escaping poverty (Dontsop et al. 2010).
These results suggest that NERICA can bring hope to
millions of poor, small-scale farmers in Africa by reducing
poverty and income inequality. Realization of this hope
requires wider dissemination of NERICA, however, which
can take place only if the seed bottlenecks and other pro-
duction constraints examined in the next section are
addressed.
LESSONS LEARNED
The use of the PVS approach allowed farmers to evaluate
the new varieties in comparison with their own material. It
enhanced capacity building and ownership of the NERICA
varieties among farmers and national research and exten-
sion communities. PVS also reduced the time involved in
the varietal release process in many countries.
Different partnerships and approaches were essential for
NERICA adoption and seed production (box 15.2). In many
countries the active involvement of government authorities
at the highest level helped stimulate uptake. Major govern-
ment initiatives on rice in Guinea, Mali, Nigeria, and
Uganda focused on NERICA varieties. In Benin a private
fertilizer and quality seed of high-yielding rice varieties,
including NERICA varieties.
The Gambia
NERICA varieties were introduced into The Gambia
through PVS trials in 1998, initially in three villages and
later more widely through farmers’ own channels and
efforts by the National Agricultural Research Institute
and the Department of Agricultural Services. The
uptake of NERICA varieties continued through the
African Rice Institute and NERICA dissemination proj-
ect funded by the African Development Bank in 2004.
Even the head of state was growing NERICA varieties
on his own land and encouraging farmers to grow these
varieties. According to FAO data, from 2001 to 2007, the
rice-growing area was relatively constant, at about
17,000 hectares. A doubling to 34,000 hectares occurred
(continued next page)
 Box 15.2 (continued)
in 2008, partly as a result of the enthusiasm generated
by NERICA varieties, the support of the government,
and the efforts of researchers and technicians. NERICA
varieties are now cultivated in all six agricultural
regions of the country. Introduction of the varieties has
brought hope for increased rice productivity and
poverty reduction.
Guinea
The first interspecific lines were introduced in Guinea
in 1997 by the Institut de Recherche Agronomique de
Guinée (IRAG) and the Service National de la Promo-
tion Rurale et de la Vulgarisation (SNPRV) through an
intensive program to introduce new upland rice vari-
eties. NERICA varieties subsequently experienced a
very rapid spread in the regions of Upper Guinea and
Forest Guinea after only two years of experimental test-
ing and PVS (Camara et al. 2002; Diagne 2006). A study
by researchers from AfricaRice, IRAG, and SNPRV in
the four regions of Guinea estimates the area under
NERICA varieties at 51,000 hectares in 2003, slightly
less than 10 percent of the 525,000 hectares used for rice
production according to Food and Agriculture Organi-
zation (FAO) statistics (FAOSTAT 2010)—a remarkable
achievement in a relatively short time. In 2007 the area
under NERICA varieties in Guinea was estimated at
82,930 hectares, 12 percent of the total rice area (ARI
2008). Estimates from the nationwide survey show that
the area under NERICA has grown significantly since
2007 to 142,391 hectares, representing about 18 percent
of the total rice area in Guinea (AfricaRice 2010).
Several factors explain the success of NERICA vari-
eties in Guinea. First, NERICA varieties appeared when
Guinea was experiencing a period of prolonged drought
that had pushed the Guinean government to seek assis-
tance from AfricaRice to put at farmers’ disposal short-
season varieties that could better cope with drought
(Diagne 2006). NERICA varieties corresponded well to
what was needed. Second, Guinean farmers were very
receptive to experimenting with new varieties in their
fields. Third, an effective partnership between national
and international institutions greatly contributed to the
rapid adoption of the new rice varieties. The World
a. Donors included the African Development Bank, the Food and Agricultural Organization, the Government of Japan, the Inter-
national Development Research Centre, the International Fund for Agricultural Development, the Rockefeller Foundation, the
United Nations Development Programme, the U.S. Agency for International Development, and the World Bank.
CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
Bank funded the program through its Special Program
for Agricultural Research in Africa (SPAAR). IRAG and
SNPRV worked together to select the most appropriate
testing sites and methods of introduction, joining hands
with an international nongovernmental organization,
Sasakawa Global 2000. This cooperation facilitated
access to inputs for farmers involved in the trials.
AfricaRice staff provided training to IRAG and SNPRV
scientists and technicians in the conduct of the PVS tri-
als and seed production throughout the duration of the
program (Kaneda 2007d; Diagne 2006).
Uganda
Rice farming is relatively marginal in East Africa, espe-
cially in Uganda. Before 2002 rice farming was practiced
mostly as a cash crop by only a few farmers on a very
small upland area and in very few small irrigated
schemes. Research support had been abandoned in the
1970s, and rice yields had gradually declined to about
0.4 ton per hectare in 2002. The introduction of NER-
ICA varieties enabled farmers to attain yields of up to
3 tons per hectare on fertile land in Uganda.
By 2002 NERICA varieties were being cultivated in
many parts of Uganda. Rice acreage increased sixfold in
the six years following the release of the new varieties,
from 6,000 hectares in 2002 to 40,000 hectares in 2008,
and the number of rice growers rose from 4,000 in 2004
to 35,000 in 2007 (Akintayo et al. 2009). Uganda
reduced its rice importation from 60,000 tons in 2005
to 35,000 tons in 2007—an almost 50 percent reduction
in rice imports, leading to savings of about $30 million
(Akintayo and others 2009). Several factors explain the
rapid uptake in Uganda. The great interest shown by the
National Institute of Agricultural Research for the new
varieties at the time of revitalizing its rice program and
the dynamism of Sasakawa Global 2000 played very
important roles in the introduction and diffusion of
NERICA varieties. The interest and dynamism of the
private sector was essential in producing and marketing
seed of NERICA varieties, as was the active involvement
of the highest political authorities of the country, par-
ticularly Uganda’s vice president (Akintayo et al. 2009).
259
 entrepreneur was instrumental in the diffusion of NERICA
varieties, paying for demonstration plots across the country
and promoting seed production (Akintayo et al. 2009).
The ARI–coordinated NERICA dissemination project
funded by the African Development Bank and numerous
smaller projects run by the Japan International Coopera-
tion Agency, FAO, and many nongovernmental organiza-
tions (NGOs) greatly accelerated the uptake of NERICA
varieties across Sub-Saharan Africa (map 15.1). Despite
these efforts, seed production remains a major bottleneck,
especially in West and Central Africa, where many farm-
ers lack access to improved rice varieties. In many coun-
tries, national seed regulatory bodies do not exist or are
dysfunctional. Functioning bodies are essential to esti-
mate and meet the demand for rice seed for different
ecosystems and consumer preferences. Incentives should
be provided to small-scale enterprises to develop viable
seed businesses.
The data on potential adoption rates of NERICA vari-
eties illustrate its tremendous potential to reach still more
farmers. The constraints to doing so are numerous and vary
from country to country. One major factor is the desperate
lack of capacity at all levels in the rice value chain, especially
the neglect of Africa’s rice research capacity. Only about
250–275 researchers (including only about 15 women) are
Map 15.1 Areas of Sub-Saharan Africa Producing NERICA Varieties, 2005 and 2006
a. 2005
PRODUCTION STATUS
(2005):
HIGH: >10,000 ha.
MEDIUM: 5,000–
10,000 ha.
LOW: <5,000 ha. 0 1,500 Km.
INTERNATIONAL BOUNDARIES 0 1,000 Mi.
Source: Akintayo et al. 2009.
260 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
involved to some extent in rice research in AfricaRice mem-
ber states, according to a 2008 survey—and most of them
work on many other crops and spend only a fraction of their
time on rice. Egypt alone accounts for the lion’s share of this
research pool, with 50 highly qualified researchers working
full time on rice, including 12 rice breeders. In comparison,
Nigeria has only 2 rice breeders.
Other issues include lack of access to seed, fertilizer, and
credit and inadequate rice production, processing, distribu-
tion, and marketing infrastructure. These factors need to be
identified, and relevant actions taken to develop the rice sec-
tor. An integrated approach from seed to plate is needed.
Africa has tremendous potential to expand the area
under rice cropping, especially in the rainfed lowland
ecosystem. It has been estimated that 200 million hectares of
rainfed lowland area could be developed for rice across the
continent. If just 1 percent of this land were allocated to
growing rice and the average yield was just 3 tons per
hectare, Africa would produce another 6 million tons of rice
a year, making a major dent in the 10 million tons of rice it
imports. The lowlands offer great potential for the sustain-
able expansion, intensification, and diversification of rice-
based systems. Given this potential, the lowland NERICA
varieties and their successors are expected to have an even
greater impact than the upland NERICA varieties.
b. 2006 IBRD 38555
MAY 2011
NERICA
DISTRIBUTION
IN AFRICA
0 1,500 Km.
INTERNATIONAL BOUNDARIES 0 1,000 Mi.
CONCLUSIONS including farmers who have never seen a new rice variety,
can profit from these new varieties, especially in rainfed
As a result of the increase in rice consumption across Sub-
ecologies.
Saharan Africa, the continent accounted for 30 percent of
AfricaRice is already working on new varieties that will
world rice imports in 2009, costing African economies an
replace the first generation of NERICA varieties, in close
estimated $4 billion. Relying on the world market to feed
collaboration with national agricultural research institutes,
Africa’s rice consumers is a risky and unsustainable strategy.
the International Rice Research Institute, and other interna-
Encouraging the development of Africa’s rice sector is cru-
tional research institutes in Japan, Europe, and the United
cial for food security. The sector could also serve as an
States. These efforts, together with a better definition of
engine for economic growth (IFPRI 2006).
farmers’ needs and consumer demands, are expected to lead
Development of Africa’s rice sector will require an inte-
to the development of varieties that will eventually outper-
grated seed to plate value chain approach. Much work is still
form the current generation of NERICA varieties.
needed to identify the varieties that respond to consumer
AfricaRice is an active member of the Coalition of
demand and grow well in the various ecosystems in which
African Rice Development, an initiative of the Japan Inter-
rice is grown in Africa, under both current and future cli-
national Cooperation Agency and the Alliance for a Green
mate scenarios. Much greater input in varietal development
Revolution in Africa, which aims to double rice produc-
is needed at the national level, and greater capacity in rice
tion in Sub-Saharan Africa between 2008 and 2018. Such
breeding is needed across the continent.
partnerships and commitment at the national and regional
NERICA varieties have shown great potential in both
level to hire, train, and retain new staff in rice research and
upland and lowland ecosystems in Africa, where they cur-
invest in rice production, storage, processing, and distri-
rently cover more than 700,000 hectares. Women farmers in
bution capacity will be instrumental in boosting Africa’s
particular have profited from NERICA varieties. More effort
rice sector.
is needed to ensure that larger numbers of rice farmers,

CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 261
 ANNEX 15.A1

Table 15.A1 Performance of NERICA Progenies and Their Parents under Various Conditions
Yield under Yield under acid
Yield under high- and drought- and and nonacid soil Susceptibility
low-input conditionsa nondrought-prone sites sites to gall midge
Item Low input High input Drought Nondrought Acid Non-acid attack
NERICA progeny line
Sample size 19 19 16 16 25 25 100
Mean (across replications and lines) 1.16 3.07 1.62 2.05 1.99 3.19 40.72
Standard deviation (across lines) 0.16 0.50 0.39 0.36 0.27 0.40 8.95
Minimum 0.91 1.70 1.08 1.49 1.57 2.15 8.76
Maximum 1.47 3.86 2.27 2.66 2.63 3.96 58.69
Sativa parent (WAB56-104)
Mean value (across replications) 0.90 3.26 1.19 1.96 1.35 3.26 52.77
T-test for equality with mean of NERICA lines
p-value 0.00 0.12 0.00 0.33 0.00 0.36 0.00
Glaberrima parent (CG14)
Mean value (across replications) 0.64 1.00 1.24 0.86 1.23 1.00
T-test for equality with mean of NERICA lines
p-value 0.00 0.00 0.00 0.00 0.00 0.00
Sativa tolerant checkb
Mean value (across replications) 1.51 1.96 12.07
T-test for equality with mean of NERICA lines
p-value 0.25 0.33 0.00
Sativa susceptible checkc
Mean value (across replications) 0.49 1.12 38.36
T-test for equality with mean of NERICA lines
p-value 0.00 0.00 0.01

Source: Authors, based on data from experiment in 1999 at AfricaRice headquarters (in M’bé) and at two of its on-farm research sites (in Man and
Korhogo). The trials used a randomized complete block design with three replicates at each site: the NERICA progenies, their two sativa and glaber-
rima parents (WAB56 104 and CG14), and a select set of sativa checks, which varied across the trials. The set of NERICA progenies also differed
across trials.
a. Yield data are mean values for two sites (M’bé and Man) from a 1999 randomized complete block design trial with three replicates at each site.
b. Moroberekan, a local sativa variety commonly used as check for drought.
c. IR20, an improved sativa variety used as check for drought.

262 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
Figure 15.A1 Performance of NERICA Progenies under Low and High Input Conditions and Major Stresses in Upland
Rice Ecosystems

a. Yield under low and high b. Yield under drought and

5 input conditions nondrought conditions
3.0

4 2.5

2.0
Yield (tones/ha)

Yield (tones/ha)
3

1.5
2
38 1.0

1
0.5

0 0
Low input use High input use Nondrought site Drought site

NERICA lines Sativa parent

Glaberrima parent Sativa tolerant check
Sativa suceptible check

c. Yield under acid and nonacid soils d. Susceptibility to African gall midge attack
4.5 70
Score for susceptibility to AfRGM attack

60
4.0
197
50
3.5
40
3.0
Yield (tones/ha)

30

2.5 20
2
21 10
2.0 1

0 102
1.5
–10
es

nt

ch

sa

1.0
ch
re
lin

al
le
nt
pa

on
ib
A

ta

pt
IC

iti
a

sis
tiv

ce

ad
ER

Sa

re

su

0.5
tr
N

er
a
tiv

tiv

th

Nonacid soil Acid soil

Sa

Sa

NERICA lines Sativa parent

Glaberrimas parent

Source: Authors, based on data from experiment in 1999 at AfricaRice headquarters (in M’bé) and at two of its on-farm research sites (in Man and
Korhogo). The trials used a randomized complete block design with three replicates at each site: the NERICA progenies, their two sativa and glaberrima
parents (WAB56 104 and CG14), and a select set of sativa checks, which varied across the trials. The set of NERICA progenies also differed across trials.
Note: Sample sizes are 19 for panel a, 16 for panel b, 25 for panel c, and 100 for panel d.

CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 263
 NOTES
1. The role of technological innovation in raising agricul-
tural productivity and fostering overall agricultural devel-
opment is documented in the World Bank’s 2008 World
Development Report.
2. The Africa Rice Center—AfricaRice for short—is the
new official name, adopted in September 2009, for the
organization previously known as the West Africa Rice Devel-
opment Association (WARDA). This intergovernmental asso-
ciation of African countries is one of the 15 international
agricultural research centers supported by the Consultative
Group on International Agricultural Research (CGIAR).
The center was created in 1971 by 11 African countries.
With the continuing expansion of its activities into Central,
East, and North Africa, the Council of Ministers passed a
resolution at their 27th meeting, held in LomÈ, Togo, in
September 2009, to change its name from WARDA to
AfricaRice. Today its membership comprises 24 countries:
Benin, Burkina Faso, Cameroon, Central African Republic,
Chad, Côte d’Ivoire, Democratic Republic of Congo, Arab
Republic of Egypt, Gabon, The Gambia, Ghana, Guinea,
Guinea-Bissau, Liberia, Madagascar, Mali, Mauritania,
Niger, Nigeria, Republic of Congo, Senegal, Sierra Leone,
Togo, and Uganda.
3. The importance of the development of NERICA vari-
eties has been widely recognized. The Consultative Group on
International Agricultural Research (CGIAR) presented the
King Baudouin Award to the Africa Rice Center (then
WARDA) in 2000 for the development of NERICA varieties.
In 2004 the center’s Dr. Monty Jones became the first African
to receive the World Food Prize, for his contribution to the
development of the family of NERICA rice varieties suitable
for upland growth conditions. Dr. Moussa SiÈ, a senior rice
breeder at AfricaRice, was awarded the prestigious Interna-
tional Koshihikari Rice Prize of Japan, in recognition of his
leadership in developing the family of NERICA rice varieties
suitable for lowland growth conditions.
4. Rice is generally associated with Asia, but it is also an
integral part of the history and culture of Africa, where it has
been grown for more than 3,000 years. Oryza glaberrima was
grown largely in the central delta of the Niger River and in
the Higher Gambia and the Casamance region of what is
now Senegal (Carney 1998; Poteres 1962, cited in Linares
2002). Oryza glaberrima was the only rice species grown in
West Africa until the 16th century, when the Portuguese
introduced the Asian species Oryza sativa in the West African
coastal regions (Linares 2002). Over the years, O. sativa grad-
ually replaced O. glaberrima, which has some unfavorable
characteristics. Farmers who still grow O. glaberrima do so
mostly for social or religious reasons, for its culinary and
nutritional qualities, or because it is better adapted to their
agroclimatic environments.
264 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
5. Seventy percent of the production increase came from
land expansion, with just 30 percent attributable to produc-
tivity enhancement.
6. Lodging is the process in which rice plants fall over at, or
just before, maturity; shattering is the process in which
plants’ panicles scatter their seeds on the ground at maturity.
7. NERICA rice varieties are interspecific but not geneti-
cally modified.
8. AfricaRice has five on-farm research sites in Côte
d’Ivoire, known as “key sites.” The sites—selected in the
early 1990s, when AfricaRice moved its headquarters to
Côte d’Ivoire from Liberia—were chosen to cover all nine
rice ecologies and the main rice-producing regions in the
forest and savanna agroecological zones of Côte d’Ivoire and
to be representative of almost all of the rice- growing ecolo-
gies in West Africa.
9. NERICA varieties have retained only about 5–10 percent
of the genome of the African parent, because of backcrosses
to the O. sativa parent. They are also based on crosses
between only a handful of O. sativa and O. glaberrima acces-
sions. AfricaRice breeders have since developed “bridge
varieties” that show full fertility when crossed with
O. glaberrima, potentially allowing the full exploitation of
all 2,300 accessions of the O. glaberrima species in the gene
bank of AfricaRice and a much greater retention of the
O. glaberrima genome. Molecular assisted breeding is also
becoming a routine tool, accelerating varietal development.
10. ARI was established in 2002 by AfricaRice and key
donor partners, including the African Development Bank,
the government of Japan, the Rockefeller Foundation, and
the United Nations Development Programme, in order to
facilitate the dissemination of NERICA varieties and other
improved varieties in Sub-Saharan Africa. This network of
rice research and extension organizations and private sector
partners plays an important role in making quality rice seed
available. This role is crucial, as seed availability is one of the
weakest links in the rice value chain in Africa.
11. Data from these national surveys, which are still being
processed, are expected to provide more accurate estimates
of the total area under NERICA in Sub-Saharan Africa. Pre-
liminary figures from these surveys and other secondary
sources suggest that about 700,000 hectares was under
NERICA in Sub-Saharan Africa in 2009.
12. The potential adoption rate estimated here is the adop-
tion rate when the entire population of rice farmers is made
aware of the existence of the NERICA varieties (Diagne and
Demont 2007). Observed adoption rates are lower than
these estimated potential adoption rates because many
farmers simply lack awareness of the NERICA’s existence.
13. PVS villages are villages in which NERICA varieties were
tested and evaluated by farmers through participatory vari-
etal selection (PVS) trials. The nearby villages are villages
not hosting PVS trials but within 2–5 kilometers radius of
the selected PVS-hosting villages.
14. With women making up 93 percent of sample farmers
in The Gambia (sample includes 39 men and 532 women),
it is not surprising that the average for the whole sample and
the sample for women are almost the same. The average
gain for men is not significant. Although seemingly small,
the 0.14 ton per hectare average gain from NERICA adop-
tion is statistically significant and represents about 15 per-
cent of the average rice yield obtained by farmers (just
under 1 ton per hectare).
15. The poor are defined using the income poverty line in
Benin, estimated at CFA 51,413 a year (about $100) in rural
areas and CFA 91,709 a year (about $180) in urban areas
(Sogbossi 2008).
REFERENCES
Adégbola, P. Y., A. Arouna, A. Diagne, and A. A. Souléïmane.
2006. “Evaluation de l’impact économique des nouvelles
variétés de riz NERICA au Bénin: évidence avec les mod-
èles basés sur l’approche contre factuel.” Paper presented
at the First Africa Rice Congress, Dar es Salaam, July
31–August 4.
Adékambi, A. S., A. Diagne, F. Kinkingninhoun-Medagbe,
and G. Biaou. 2009. “The Impact of Agricultural Tech-
nology Adoption on Poverty: A Case Study of NERICA
Varieties in Benin.” Paper presented at the meetings of
the International Association of Agricultural Econo-
mists, Beijing, August 16–22.
AfricaRice. 2010. Rice Data Systems for Sub-Saharan Africa:
Contribution to the Japan–AfricaRice Emergency Rice
Project. Updated synthesis report submitted to the gov-
ernment of Japan, September 30, Cotonou.
Afriquenligne 2009. “Le riz NERICA pour assurer la sécurité
alimentaire.” In Actualités africaine en ligne. http://www
.afriquenligne.fr/actualites/economie/le-riz-nerica-
pour-assurer-la-securite-alimentaire-200804252475
.html.
Agboh-Noameshie, A. R., F. M. Kinkingninhoun-Medagbe,
and A. Diagne. 2007. “Gendered Impact of NERICA
Adoption on Farmers’ Production and Income in Cen-
tral Benin.” Paper presented at the meetings of the
African Association of Agricultural Economists, August
20–22, Accra.
Akintayo, I., O. Ajayi, R. A. Agboh-Noameshie, and B. Cissé.
2009. Report of the Seventh African Rice Initiative Steering
Committee/Experts Meeting, May 4–6, Cotonou.
ARI (African Rice Institute). 2008: Initiative africaine sur le
riz: rapport annuel 2008. Centre du Riz pour l’Afrique
ADRAO, Cotonou.
CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
Bentley, J., C. Almekinders, D. Yawovi, and D. Dalohoun.
2008. Center Commissioned External Review of the
African Rice Initiative. AfricaRice, Cotonou.
Bergman-Lodin, J. 2005. “The NERICA Conundrum: From
Rice to Riches.” Lund University, Department of Eco-
nomic and Social Geography, Lund, Sweden.
Camara, B., A. Conde, M. S. Diallo, M. B. Barry, B. Sekou,
and T. Berhe. 2002. NERICA, the Guinea Experience.
Report of the Departments of National Research, Exten-
sion and Rural Development and Sasakawa Global
2000/Guinea Agricultural Project, Conakry.
Carney, J. A. 1998. “The Role of African Rice and Slaves in
the History of Rice Cultivation in the Americas.” Human
Ecology 26 (4): 525–45.
Dalton, Timothy. J. 2004. “A Household Hedonic Model of
Rice Traits: Economic Values from Farmers in West
Africa.” Agricultural Economics 31: 149–59.
Diagne, A. 2006. “Le NERICA: mythe et réalités?” Grain de
sel, la revue de l’inter-réseaux (March–August): 34–37.
———. 2009. “Technological Change in Smallholder Agri-
culture: Bridging the Adoption Gap by Understanding Its
Source.” Paper presented at the conference on Agricul-
ture for Development sponsored by the African Eco-
nomic Research Consortium, Mombasa, May 28–29.
Diagne, A., and M. Demont. 2007. “Taking a New Look at
Empirical Models of Adoption: Average Treatment Effect
Estimation of Adoption Rates and Their Determinants.”
Agricultural Economics 37 (2–3): 201–10.
Diagne, A., G. S. Midingoyi, and F. Kinkingninhoun. 2009.
“The Impact of NERICA Adoption on Rice Yield in West
Africa: Evidence from Four Countries.” Paper presented
at the meetings of the International Association of Agri-
cultural Economists Conference, Beijing, August 16–22.
Diagne A., M. J. Sogbossi, S. Diawara, and Camara, A. 2006.
L’étendu du succès de la dissémination des variétés NER-
ICA en Guinée: estimation des superficies emblavées. West
Africa Rice Development Association, Cotonou.
Dibba, L. 2010. “Estimation of NERICA Adoption Rates and
Impact on Productivity and Poverty of Small-Scale Rice
Farmers in The Gambia.” M.Phil. thesis, Kwame
Nkrumah University, Faculty of Agriculture, Depart-
ment of Agricultural Economics, Agri-business and
Extension, Kumassi, Ghana.
Dibba, L., A. Diagne, and S. Fialor. 2010. “Impact of NER-
ICA Adoption on Productivity and Poverty of the Small-
Scale Rice Farmers in The Gambia.” Kwami Nkrumah
University of Science and Technology Kumasi Ghana and
Africa Rice Center, Cotonou.
Dibba, L, A. Diagne, F. Simtowe, M.J. Sogbossi and M.
Mendy. 2008a. “Diffusion and Adoption of NERICA in
Gambia.” Africa Rice Center, Cotonou.
265
 ———. 2008b “The Impact of NERICA Adoption on Yield
in Gambia. Africa Rice Center Cotonou.
Dontsop, P., A. Diagne, V. O. Okuruwa, and V. Ojehomon.
2010. “Impact of Improved Rice Technology on Income
and Poverty among Rice Farmers in Nigeria: The Case of
NERICA Varieties.” University of Ibadan Nigeria and
Africa Rice Center, Ibadan.
FAO (Food and Agriculture Organization). 2009. “The Spe-
cial Challenges for Sub-Saharan Africa.” High-Level
Experts’ Forum, Rome, October 12–13.
———. 2010. Food and Agricultural Organization statis-
tics. http://www.fao.org/corp/statistics/en/.
Glove, S. 2009. “Agricultural Technology Dissemination in
West Africa: A Quantitative Analysis of NERICA Adop-
tion and Its Impact in Guinea and The Gambia.” MSc.
thesis, University of London, Department of Develop-
ment Economics, the School of Oriental and African
Studies.
Gridley, H., and M. Sié. 2008. “The Role of PVS: Farmer in
the Driving Seat.” In NERICA: New Rice for Africa. A
Compendium, ed. E. Somado, R. Guei, and S. O. Keya.
AfricaRice, Cotonou.
IFPRI (International Food Policy Research Institute), in col-
laboration with IITA (International Institute of Tropical
Agriculture), CORAF/WECARD (Conseil Ouest et Cen-
tre Africain pour la Recherche et le Développement Agri-
coles/West and Central African Council for Agricultural
Research and Development), and ECOWAS (Economic
Community Of West African States). 2006. Regional
Strategic Alternatives for Agriculture-Led Growth and
Poverty Reduction in West Africa. Final Draft Report.
Washington, DC.
IRAT (Institut de Recherche Agronomique Tropicale). 1967.
“Les variétés de riz au cercle de Banfora. IRAT/Haute
Volta.” L’Agronomie Tropical 24: 691–707.
Johnson, D. E 1997. Weeds of Rice in West Africa. West Africa
Rice Development Association (WARDA). Bouaké, Côte
d’Ivoire.
Jones, M. P., M. Dingkuhn, G. K. Aluko, and M. Semon.
1997. “Interspecific Oryza sativa × O. glaberrima Steud.
Progenies in Upland Rice Improvement.” Euphytica 92:
237–46.
Kaneda, C. 2007a. “Breeding and Disseminating Efforts of
NERICA: Breeding of Upland Rice.” Journal of Japanese
Tropical Agriculture 51 (1): 1–4.
———. 2007b. “Breeding and Disseminating Efforts of
NERICA: Evaluation of Important Characteristics.”
Journal of Japanese Tropical Agriculture 51 (2): 41–45.
———. 2007c. “Breeding and Disseminating Efforts
of NERICA: Important Characteristics of Lowland
266 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA)
Rice.” Journal of Japanese Tropical Agriculture 51 (3):
79–83.
———. 2007d. “Breeding and Disseminating Efforts of
NERICA: Efforts for Dissemination of NERICA in
Africa.” Journal of Japanese Tropical Agriculture (51): 4:
145–51.
Kijima, Y., K. Otsuka, and D. Sserunkuuma. 2008. “Assessing
the Impact of NERICA on Income and Poverty in Cen-
tral and Western Uganda.” Agricultural Economics 38:
327–337.
———. 2009. “Determinants of Changing Behaviors of
NERICA Adoption: An Analysis of Panel Data from
Uganda.” University of Tsukuba, Graduate School of Sys-
tems and Information Engineering, Tsukuba, Ibaraki,
Japan.
Kijima, Y., D. Sserunkuuma, and K. Otsuka. 2006. “How
Revolutionary Is the ‘NERICA Revolution’? Evidence
from Uganda.” Developing Economies 44 (2): 252–67.
Linares, O. F. 2002. African Rice Oryza Glaberrima: History
and Future Potential. Smithsonian Tropical Research
Institute, Balboa-Ancon, Panama.
Ndjiondjop, M-N., K. Semagn, M. Cissoko, M. P. Jones, and
S. McCouch. 2008. “Molecular Characterization of NER-
ICA Lines.” In NERICA: New Rice for Africa. A Com-
pendium, ed. E. Somado, R. Guei, and S. O. Keya. Coto-
nou: AfricaRice.
Porteres, R. 1962. “Berceaux agricoles primaires sur le con-
tinent africain.” Journal of African History 3 (2): 195–210.
Science Council. 2007. Report of the Fifth External Program
and Management (EPMR) of the Africa Rice Center
(WARDA). Consultative Group on International Agri-
cultural Research (CGIAR), Science Council Secretariat,
Rome.
Seck P. A., E. Tollens, M. C. S. Wopereis, A. Diagne, and I.
Bamba. 2010. “Rising Trends and Variability of Rice
Prices: Threats and Opportunities for Sub-Saharan
Africa.” Food Policy 35 (5): 403–11.
Sogbossi, M-J. 2008. “Riziculture et réduction de la pauvreté
au Bénin: impact de l’adoption des nouvelles variétés riz-
icoles sur la pauvreté et l’inégalité du revenu dans le
Département des Collines Bénin.” Master’s thesis, Univer-
sity of Abomey-Calavi, Département d’Economie Socio-
anthropologie et Communication, Abomey-Calavi, Benin.
Solidarité et Progrès. 2008. “Mauvais temps pour les
malthusiens! La production de riz explose en Afrique.”
http://www.solidariteetprogres.org/article5783.html.
Somado, E., R. Guei, and S. O. Keya. 2008. NERICA: New
Rice for Africa. A Compendium. Cotonou: AfricaRice.
Spencer, D., A. Dorward , G. Abalu, D. Philip, and D. Ogung-
bile. 2006. Evaluation of Adoption of NERICA and Other
 Improved Upland Rice Varieties Following Varietal Promo-
tion Activities in Nigeria. Reported prepared for the
Gatsby and Rockefeller Foundations, New York.
Wopereis, M. C. S., A. Diagne, J. Rodenburg, M. Sié, and
E. A. Somado. 2008. “Why NERICA Is a Successful
Innovation for African Farmers: A Response to Orr et
al. from the Africa Rice Center.” Outlook on Agriculture
37 (3): 169–76.
World Bank. 2008. World Development Report 2008: Agricul-
ture for Development. Washington, DC: World Bank.

CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 267
 CHAPTER 16
Fertilizer in Kenya: Factors Driving the
Increase in Usage by Smallholder Farmers
Joshua Ariga and T. S. Jayne
ertilizer use is notably lower in most of Sub-Saharan
F
Africa than in other developing regions. Too little
irrigation and varieties unresponsive to fertilizer may
explain this low use to some degree. But more often, the
causes are lack of credit, long distances between farmers and
the nearest fertilizer retailer, weak market infrastructure, and
lack of government support. Indeed, in many countries, the
withdrawal of state input delivery systems has led to a reduc-
tion in fertilizer use as commercial distribution systems
compete with subsidized government programs.
Kenya, however, stands as a notable departure from this
common Sub-Saharan African narrative. In the early 1990s
fertilizer markets were liberalized, government price con-
trols and import licensing quotas were eliminated, and fer-
tilizer donations by external donor agencies were phased
out. Fertilizer use then almost doubled over the 15-year
period from 1992 to 2007, with much of the increase attrib-
utable to smallholder farmers. In the productive farming
areas of western Kenya, rates of fertilizer application on
maize are comparable with rates in Asia and Latin America.
Kenya’s economy is predominantly agrarian, with more
than 70 percent of its people dependent on agriculture-
related farm and off-farm activities for their livelihoods
(Ministry of Agriculture 2004). Food security is a concern
for the 60 percent of the population living below the $1-a-
day poverty line. While increasing the available supply of
Joshua Ariga is a senior research fellow with the Tegemeo Institute of Agricultural Policy and Development, Egerton University, and a research specialist at
Michigan State University. T. S. Jayne is Professor, International Development, Michigan State University.
food is an important goal, the problem of access to food is
perhaps the primary cause of food insecurity in Kenya.
Ensuring access to food requires that the poor are able either
to produce or to buy enough food for a healthy diet.
Because of unpredictable weather and poor infrastruc-
ture, producers and consumers of agricultural products in
Kenya face volatile market prices, with periods of surplus
production providing a boon for consumers and periods of
deficit benefiting a relatively few producers who are net sell-
ers and hurting the majority of consumers. The justification
for state participation in the market for agricultural prod-
ucts in Kenya has been to maintain prices at levels that both
provide incentives to raise farm incomes for producers and
ensure that consumers can access food, the classic “food
price dilemma.”
Before 1990 Kenya addressed this dilemma through direct
participation in input and output markets for national
“strategic” crops through either state-run agencies that set
prices at panterritorial levels or ostensible farmer organiza-
tions that were managed by state-connected political agents
or their surrogates. For example, in the coffee sector, the gov-
ernment helped enact laws that created the Coffee Board of
Kenya and Kenya Planters Cooperative Union; for
pyrethrum (chrysanthemums), it encouraged the creation of
the Pyrethrum Board of Kenya; and for milk, tea, and maize,
the government helped create the Kenya Cooperative
269
 Creameries, the Kenya Tea Development Agency, and the
National Cereals and Produce Board (NCPB). In its heyday,
the NCPB generally bought maize grain from farmers at
higher-than-market prices and sold it to industrial maize
millers at below-market prices. On the input side, the 1970s
and 1980s brought the formation of the state-run Kenya
National Trading Corporation (KNTC) and the Kenya
Grain Growers Cooperative Union (KGGCU), later the
Kenya Farmers Association (KFA), which worked with the
output organizations.
Of all the crops mentioned in the previous paragraph,
policy makers in Kenya—and indeed, throughout East
Africa—have been most concerned with increasing fertilizer
use on maize, the main food security crop in the region. For
a number of reasons, however, state efforts in the 1980s to
improve food security through increased production and
incomes did not produce desired results. As a result, several
reform measures were implemented that sought to achieve
food security objectives in a more efficient way, in the lines
of a laissez faire or competitive markets dogma.
Fertilizer and maize market reforms
The period before market reforms in Kenya was character-
ized by a predictable pattern involving the participation of
state-run agencies or private farmer organizations (with
heavy state intervention in their management) in input and
output markets for import and export, distribution, and
retailing. Although these state agencies continually rein-
vented themselves under different names, particularly when
they came under scrutiny for corruption and unsustainable
budgets, their reincarnation followed the same general
modus operandi, and all eventually failed to achieve their
goal of improving the livelihoods of smallholder farmers.
To put the reform process into perspective, it is impor-
tant to recognize that agricultural policy in Kenya has
also gone through a number of key phases characterized
by an unpredictable shelf life. In the immediate postinde-
pendence period (late 1960s), agricultural policy was
concerned with supporting a smooth transfer of prime
land from white settlers to indigenous Kenyans with help
from state-supported agencies in the production and mar-
keting of produce (such as NCPB for maize). Agricultural
inputs were marketed through the farmers’ union, KFA,
and credit was provided through the Agricultural Finance
Corporation (AFC).
In the 1960s KFA could, for instance, offer inputs on
credit (through AFC) to select farmers, who repaid the
union after harvesting the crop and delivering it to relevant
270 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
marketing agencies such as the NCPB (which deducted the
cost of the loans on behalf of the KFA and AFC). To deal
with high prices and a weak distribution network for small-
holder farmers, fertilizer subsidies were also introduced
through these agencies (Ariga, Jayne, and Nyoro 2006). This
conflict of interest across interlinked agencies generated
widespread corruption and bureaucratic costs that led to a
policy change in 1972 in favor of introducing another
agency (KNTC), which was tasked with importing fertilizer,
while the KFA was to be the distributor. This shift was
intended to increase competition within Kenya’s market for
fertilizer, but it did not succeed in keeping fertilizer prices at
low levels, and the agencies, influenced by the state, fell into
the same patterns of bureaucracy and corruption as before.
On the output side, the NCPB controlled maize prices at all
levels of the market chain (Nyoro, Kiiru, and Jayne 1999). By
setting fixed panterritorial prices and removing arbitrage
opportunities for all market participants, these entities sti-
fled private trade. Requirements that private traders apply
for permits to transport grain across district boundaries
made the situation more difficult.
In the 1980s the government started relaxing its monop-
oly, allowing the private sector to compete with state agen-
cies, albeit under state rules. Fertilizer traders were to adhere
to official prices, and the state influenced competition
through strict trade licensing requirements and control of
the allocation of scarce foreign exchange to importers (Arg-
wings-Kodhek 1996). Licensing and allocation of foreign
exchange provided rent-seeking opportunities for public
sector officials (Kimuyu 1994). While the controlled pricing
structure was designed to improve farmers’ access to fertil-
izer, it had the opposite effect in geographically remote
areas, where the controlled prices were too low for fertilizer
retailers to recoup the costs of transporting fertilizer from
district towns to remote areas. Hence, retailers in remote
locations were less likely to stock fertilizer than those in
more urbanized areas, leaving the average distance traveled
by farmers to procure fertilizer relatively high in the 1980s
and early 1990s.
In addition to the market inefficiencies they created, state
agencies also imposed a heavy burden on public resources in
Kenya, contributing to deficits and inflation in the 1980s. A
decline in budgetary support to the agricultural sector by
late 1980s probably contributed to the subsequent decline in
agricultural growth, as did the mismanagement of agricul-
tural institutions, the ad hoc reform agenda, withholding of
donor funds over disagreements about democracy and gov-
ernance, and depreciation of the Kenyan shilling, the last of
which raised input prices (Argwings-Kodhek 2004). In the
late 1980s and early 1990s, however, Kenya began easing
trade restrictions in the fertilizer and maize markets. The
government started removing some import quota restric-
tions, for example, in January 1990 and abolished licensing
requirements for fertilizer imports in 1992.
In a major policy change, the government liberalized the
fertilizer subsector specifically in 1993 to allow the partici-
pation of the private sector in importing, local trading, and
distribution of fertilizer. Coupled with the liberalization of
the foreign exchange regime in 1992, these changes in the
policy environment led to the entry of a significant number
of private sector firms in importing, wholesaling, distribu-
tion, and retailing of fertilizer (Wanzala 2001). Government
price controls and import licensing quotas were ultimately
eliminated, and fertilizer donations by external donor agen-
cies were phased out. Maize trading controls were relaxed in
the early 1990s to allow private traders to transport a few
bags across districts with permission from government offi-
cials, a situation that led to rent-seeking behavior and
increased costs for businesses (Kimuyu 1994). The NCPB,
however, helped stabilize prices for producers and con-
sumers by continuing to buy maize (mostly from large pro-
ducers) at above-market prices and, during shortages, sell-
ing it to consumers at subsidized prices (Jayne, Myers, and
Nyoro 2008).
With the participation of stakeholders from all facets of
society in the 1990s and 2000s, a number of government
policy papers emphasized a multisectoral approach to rural
development in Kenya, including private-public synergies in
development. By 1996, 12 major importers, 500 wholesalers,
and roughly 5,000 retailers were distributing fertilizer in
Kenya (Allgood and Kilungo 1996), and by 2000 the num-
ber of retailers was estimated to have risen to between 7,000
and 8,000 (IFDC 2001). (These are estimates because there
is no comprehensive business registry or database covering
all types of businesses in Kenya.) Even with the easing of
trade restrictions, the high costs of upland transportation
and logistical problems at the port of Mombasa continued
to inflate the cost of fertilizer and reduce effective demand
(Wanzala 2001; Ariga and Jayne 2008). Although markups
are less than 11 percent of the farm gate price of fertilizer in
western Kenya, in a number of farming areas no fertilizer is
applied because of the risk from markets, poor rainfall, and
agricultural conditions. But in other areas, application rates
rival those of Asia, and selling fertilizer is clearly profitable.
The overall trend in national consumption of fertilizer has
followed a steady growth path since 1990, with government
imports declining and the role of the private sector increas-
ing (figure 16.1).
Since 2007 a major escalation in the world price of fertil-
izer has led to increased government involvement in fertil-
izer marketing. The post-2007 period, which has been
marked by uncertain policy regimes, follows a fairly stable
and transparent period since 1993. In Kenya, in a move to

Figure 16.1 Trends in Consumption, Commercial Imports, and Donor Imports of Fertilizer in Kenya, 1990/91–2010/11

600

500
Thousands of metric tons

400

300

200

100

0
9

1
1

/9

/0

/0

/0

/0

/0

/1
/9

/9

/9

/9

98

00

02

04

06

08

10
90

92

94

96

19

20

20

20

20

20

20
19

19

19

19

Imports Consumption Donor / state imports Projected

Source: Estimated from Kenya Ministry of Agriculture data by the authors.

Note: In 2004 and 2008, respectively, NCPB imported approximately one-third and 40 percent of national fertilizer needs. The estimate given for 2010/11
is a projection that includes private and government imports. The shaded years cover the time period after 2006/07, when government imports/subsidies
were reinstated partly as a reaction to deficits in maize production and postelection disruptions of agricultural activities.

CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 271
 bolster agricultural production after postelection violence
led to disruption of farm activities, NCPB imported fertil-
izer in 2008 but delivered it to farmers late, which con-
tributed to the low levels of maize production that year. This
situation, in turn, created pressure from some farmer lobby
groups and activists for increased subsidization of inputs
(fertilizer and seed) to raise productivity of maize to
counter an expected increase in hunger in 2009. In 2009 the
government of Kenya imported substantial amounts of fer-
tilizer through NCPB to be distributed through its branches
and select private retailers at prices that included a 40 per-
cent subsidy. In early 2010 news reports indicated that the
government planned to import 1.5 million bags (75,000
tons) of fertilizer.1 Table 16.1 details key points in the liber-
alization of the maize sector, especially the role played by
NCPB, the main grain-marketing state agency.
Although the increased participation of the Kenyan gov-
ernment in the agricultural market was expected to be
short-lived and not significant enough to disrupt private
sector investments, unforeseen events (such as poor rains in
recent years) may mean that occasional state subsidies will
be implemented over the upcoming few years in an attempt
to meet national food requirements, particularly as political
pressure for such measures increases. If recent indications
are reliable, the incidence of government subsidies will
probably decline after a few years as they become unsus-
tainable unless international partners shoulder some of the
responsibility.

Table 16.1 Evolution of Maize and Fertilizer Market Policy Reforms in Kenya, 1988–2010
State marketing agency
1988: NCPB faces deficits and is financially
restructured. NCPB depots are phased
out. NCPB debts are written off; crop
purchase fund is established but not
replenished.
Early 1990s: NCPB narrows its margins.
Private sector finds it unprofitable to
reach remote areas.
1995: Donor pressure leads to NCPB
being restricted to the role of buyer and
seller of last resort. NCPB market share
declines to 10–20 percent of marketed
maize trade. NCPB operations are
confined mainly to high-potential areas
of western Kenya.
2000 onward: NCPB is provided with funds 1997–2005: External trade and tariff rate levels
to purchase a greater volume of maize.
NCPB’s share of total maize trade rises
to 25–35 percent of total marketed maize.
2005 onward: The government withdraws
Maize market policy Fertilizer market policy
1988: Cereal Sector Reform Program envisages Pre-1990: KGGCU / KFA and KNTC are
widening of NCPB price margin. In fact, main input agencies. Mismanagement and
margin narrows. Proportion of grain that deficits are common. Government control
millers are obliged to buy from NCPB of the fertilizer industry is heavy. Imports
declines. Limited unlicensed maize trade is are poorly coordinated, leading to
allowed. State sets all prices for maize surplus/deficits. Licensing of private trade in
grain and flour. the late 1980s is controlled but under
panterritorial pricing. State agencies are
financially weak.
1991: Local and international pressure for
reforms builds up. Interdistrict trade is
further relaxed.
1992: Kenya moves from a one-party political 1992: Foreign exchange regime is liberalized.
system to a multiparty system. Restrictions Fertilizer import restrictions are relaxed.
on maize trade across districts are reimposed. 1993: Fertilizer market is liberalized.
NCPB is unable to defend ceiling prices. Private traders are allowed to import and
1993: Maize and maize meal prices deregulated. distribute. State and donor imports decline
Import tariffs are abolished. No subsidies are dramatically.
provided to registered millers. 1994: Custom duty and value added tax
(VAT) are removed.
1995: Internal maize and maize meal trade is 1996: Number of market entrants is
fully liberalized. Maize import tariff of 30 estimated at 12 major importers, 500
percent is reimposed. 1996: Export ban is wholesalers, and roughly 5,000 retailers
imposed after poor harvest. (Allgood and Kilungo 1996)
1997: Import tariff is imposed after poor
harvest.
change frequently and become difficult to
predict. NCPB producer prices are normally
set above import parity levels
the import tariff on maize entering Kenya from
East African Community member countries. An
official 2.75 percent duty is
still assessed, and a variable import duty is
still assessed on maize entering Kenya
through the Mombasa port.
(continued next page)

272 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
 Table 16.1 (continued)
State marketing agency
2008: World food prices are high. NCPB is 2008: Postelection violence erupts. African
asked to sell subsidized grain to millers,
who in turn could lower prices for
consumers. State has difficulty enforcing
and monitoring at millers’ end because of
unknown milling costs. Allegations of
corruption emerge.
2009: Briefcase firms and NCPB employees 2009: Imports continue, but domestic maize
take advantage of crisis and subsidy
arrangements to favor some firms in
return for kickbacks. Weaknesses in
disaster preparedness, institutions, and
food policy are revealed. Top management
of NCPB and some Ministry of Agriculture
officials are fired for corruption during
the crisis.
2010: NCPB allocates funds to buy maize
produced in eastern Kenya during short
rainy season.
Maize market policy
Centre for Open Governance estimates that
3.5 million bags of maize are destroyed.
NCPB begins importing maize from United
States and South Africa in late 2008. An
estimated 5 million bags arrive, according
to the center.
production is greater than expected. Claims
of monopoly at Mombasa port involving grain
handling (particularly related to one large grain distributed through private trade networks.
handler, Grain Bulk Handlers Limited) and
milling arise but are not substantiated.
2010: Short rainy season results in large
amount of maize production, but farmers
report poor prices from private traders.
Fertilizer market policy
2008: High world prices for fertilizer
exacerbate food crisis effects from
postelection violence. Prices more than
double. Petrol and transport costs also
go up.
2009: NCPB imports state- subsidized
fertilizer to aid in recovery from
postelection violence. Fertilizer is
2010: State imports more than 30,000 tons
of fertilizer and distributes to vulnerable
farmers. Distribution is managed by NGOs.

Source: Authors, adapted from Ariga and Jayne (2008).

THE EFFECT OF REFORMS IN FERTILIZER
AND MAIZE MARKETS
Data collected in household surveys and found in secondary
sources can be used to analyze the effects of policy reforms
on fertilizer and maize markets in Kenya during the 1980s,
1990s, and 2000s. The panel data consist of nationwide rural
household panel survey data covering the 1996/97,
1999/2000, 2003/04, and 2006/07 crop seasons. The panel
household survey was designed and implemented under the
Tegemeo Agricultural Monitoring and Policy Analysis Proj-
ect (TAMPA), implemented by Egerton University/Tegemeo
Institute with support from Michigan State University. Out
of the national sample of 1,260 households, a balanced
panel of 899 households was interviewed in all four periods.
Other data, such as monthly maize price levels and NCPB
maize purchases and sales, were obtained from various
Kenyan government ministries.
For analytical convenience, the survey sample has been
classified into zones based on agro-ecological characteristics,
districts, and agricultural production potential. Each of
these agro-ecological zones has been split into two broader
categories—high-potential and low-potential regions—based
on soil quality, rainfall, yield potential, and fertilizer use.2
Processes leading to growth in smallholder
fertilizer use and maize yields
Figure 16.2 provides a schematic description of how public
investments in market infrastructure and policy reform
within fertilizer and maize markets generated specific
responses from the private sector and smallholder farmers.
The basic story is one of synergies between liberalization of
input and maize markets and public investments in support
of smallholder agriculture that led to substantial private sec-
tor investment in fertilizer retailing and maize marketing,
which in turn resulted in an impressive increase in the use
of fertilizer by smallholder farmers and in the maize yields
of smallholder farms over 1997–2007.
Distance from farm to fertilizer seller
One indicator of how reforms have contributed to fertilizer
availability in Kenya is the distance farmers now have to
travel to buy fertilizer compared with the distance in the
prereform period (figure 16.3). This variable is a measure of
increased private sector competition leading to more invest-
ment in locating retail services closer to producers, with
retailers opening stores in new catchment areas that were
hitherto not serviced by the government-run input system
and opening additional stores in already-served areas to
capture more business.
In general, distances in the low-potential region are
longer than those in the high-potential region, which is
one reason why the sample is split into two broad groups
in the analysis presented here. The private sector invest-
ment in fertilizer trade has expanded rapidly after the state
allowed competition and removed trade restrictions.
Although absolute distances are generally higher for the

CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 273
 Figure 16.2 Synergies Increasing Fertilizer Use and Maize Yields by Smallholder Farmers

Public investments:
1. Major investment in rural feeder roads
2. Generation and release of new maize varieties by Kenya Agricultural Research
Institute and by private seed firms

Policy reforms—fertilizer marketing: Policy reforms—maize marketing:

1. Price controls on fertilizer abolished
2. Private fertilizer trade fully legalized
3. Fertilizer import quotas eliminated
4. Government auctioning of free donor
fertilizer phased out; no competing
fertilizer subsidy program (1990–2007)
1. Barriers to private maize marketing
eliminated by 1995
2. Maize meal price controls eliminated
in 1993
3. Closing of buying stations in most parts
of the country; stations remain active in
only three or four surplus
maize-producing districts

Private sector responses:

1. Rapid expansion in private fertilizer wholesaling and retailing, reducing the
distance farmers travel to nearest fertilizer retailer
2. Reduction in fertilizer marketing costs observed between off-loading at Mombasa
port and farm gate level
3. Reduction in distance traveled by farmers to point of maize sale to private trader
4. Increase over time in maize/fertilizer price ratios

Smallholder farmer responses:

1. Rise in the percent of farmers using fertilizer and hybrid maize seed
2. Increase in maize yield and maize production
3. Increase in percent of farmers selling maize

Source: Authors.

Figure 16.3 Average Distance from Farm to Fertilizer low-potential region, the rate at which distances have
Seller declined is generally higher than in the high-potential
region. The consumption of fertilizer has not followed the
15.0 same regional pattern, however, implying the presence of
Distance to fertilizer seller (kilometers)

15
other constraints.
10.6
10 Increasing proportion of households using fertilizer
6.9 Data from a balanced panel of households in Kenya show
5 4.3 4.1
that the percentage of households using fertilizer on at least
3.0
2.3
2.9 one farm plot rose from 59 percent in 1997 to 72 percent in
2007 for the national sample (table 16.2). However, there
0 are differences in growth across agro-ecological zones and
97

00

04

07

97

00

04

07

the two broad regions of interest. For the high-potential

19

20

20

20

19

20

20

20

High-potential region
Source: Authors, estimated from Tegemeo Institute/Egerton University
household surveys, 1997, 2000, 2004, and 2007.
Note: The high-potential region includes the districts of Bomet, Bungoma,
Gishu, Kakamega, Kisii, Meru, Muranga, Narok, Nakuru, Nyeri, Trans
Nzoia, Uasin, and Vihiga. The low-potential region includes the districts
of Kilifi, Kisumu, Kitui, Kwale, Machakos, Makueni, Mwingi, Siaya, Taita,
and Taveta.
Low-potential region
region, which began with a much higher proportion of
fertilizer users, use increased from 77 percent to 91 per-
cent of households between 1997 and 2007.Use in the
low-potential region more than doubled during the same
period, increasing from 12 percent in 1997 to 26 percent
in 2007, but the proportion of households using fertilizer
remained relatively small.

274 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
 Table 16.2 Share of Maize-Growing Households Using Fertilizer, by Region and Zone
(percent)
Region and zone 1997 2000 2004 2007
High-potential region
Western transitional 41 65 71 81
High-potential maize zone 84 89 89 92
Western highlands 78 90 91 95
Central highlands 90 91 91 93
Subtotal 77 85 86 91
Low-potential region
Coastal lowland 4 4 5 11
Eastern lowland 26 27 47 48
Western lowland 2 5 7 13
Subtotal 12 14 23 26
Grand total 59 65 68 72

Source: Authors, estimated from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007.
Note: This sample consists of a balanced panel of 899 households interviewed in all four periods.

Dynamics of fertilizer application rates Figure 16.4 Average Difference in Mean Household
Fertilizer Application Rates from 1997
By examining fertilizer application rates for households
Levels, by Region
using fertilizer and for the whole sample of households sur-
veyed, including nonusers, one can observe a trend in the 20 19.0
intensity of fertilizer use (kilogram per acre of maize) since
Differences in means
(kilograms per acre)

the inception of fertilizer reforms in the 1990s. For compar- 15

ison at the household level, plot-level application rates using
10.7
plot area (acres) as weights are aggregated.3 First, for each 10
region, the differences between the weighted mean house- 7.8
hold application rates (kilograms per acre) for the years 5
2000, 2004, and 2007 are compared with the rates for 1997, 1.5 1.9
revealing whether application rates increased, decreased or 0.1
0
remained the same in subsequent years compared with the
00

04

07

00

04

07
20

20

20

20

20

base period of 1997. Figure 16.4 shows the trend in this
indicator using the base year 1997.
Clearly, weighted mean household application rates
have increased relative to those in 1997 (and also from
one year to the next) for each region. It is also obvious
that in absolute terms, the differences are much smaller in
the low-potential region, probably as a result of lower
application rates and the lower proportion of fertilizer
users (the latter shown in table 16.2). Table 16.3 supplies
information on fertilizer application rates (kilograms per
acre) for two sets of households: the whole sample of
households surveyed (the sum of fertilizer users and
nonusers), and a subset consisting of households that
used fertilizer (only users).
As shown in table 16.3, fertilizer application rates have
increased from their 1997 levels for all regions and agro-
zones, although intensities differ across regions and agro-
zones. The high- potential maize zone (HPMZ), which
20
High-potential region Low-potential region
Source: Authors, estimated from Tegemeo Institute/Egerton University
household surveys, 1997, 2000, 2004, and 2007.
includes Trans Nzoia and Uasin Gishu, and the Central
Highland zone have some of the highest application rates in
the sample. For example, the HPMZ’s rates for fertilizer
users only in 1997 and 2007 are 157.3 and 181.1 kilograms
per hectare, similar to rates in Asia, which benefited from
the green revolution and now has one of the highest fertil-
izer application rates in the world.
Trend in maize yields for unfertilized plots and
different seed technologies
As shown in figure 16.4 and table 16.3, fertilizer application
has increased significantly since the reforms of the 1990s,

CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 275
 Table 16.3 Fertilizer Application Rates for Maize-Growing Households in Kenya, by Region and Zone
(kilograms per acre)
Sum of fertilizer users and nonusers Fertilizer users only
Region/zone 1997 2000 2004 2007 1997 2000 2004 2007
High-potential region
Western transitional 23.0 47.1 46.5 57.3 57.5 73.0 63.8 71.8
High-potential maize zone 53.1 58.5 60.9 65.4 63.7 66.5 70.4 73.3
Western highlands 26.9 40.6 49.6 48.4 36.3 45.4 54.0 51.7
Central highlands 62.2 68.4 73.4 67.2 68.8 77.9 84.2 74.1
Subsample 46.3 56.1 59.5 61.6 60.6 66.7 70.0 69.6
Low-potential region
Coastal lowland 0.4 0.8 0.1 1.6 10.4 19.6 2.1 13.9
Eastern lowland 3.1 5.7 8.3 9.6 12.1 24.8 19.7 23.9
Western lowland 0.4 0.5 0.9 2.3 21.3 16.4 19.4 18.6
Subsample 1.4 2.8 4.0 5.3 12.7 23.6 19.1 22.0
Grand total 33.3 38.0 43.1 45.0 58.0 63.8 65.2 64.7

Source: Authors, estimated from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007.
Note: When estimating rates for the whole set of households, “zeros” are used for households not using fertilizer, making these rates lower or equal
to rates calculated for the group including fertilizer users only (depending on presence of nonusers).

when markets were opened up to competition. But have the
reforms and the rise in consumption of fertilizer, had an
effect on maize production? Looking at the dynamics of
maize yields for plots that received fertilizer and those that
did not across the survey years and for different seed types
provides information on this question.
First, as shown in figure 16.5, yields from fertilized and
unfertilized plots generally increased between 1997 and
2007. Second, irrespective of fertilization, yields for hybrid
seed plots are higher than those for nonhybrid plots. Third,
for each seed technology, yields for fertilized fields are
higher than those for unfertilized fields. And finally, yields
for plots that receive fertilizer and use hybrid seeds are the
highest for each of the four survey years.4
Although yields generally increased between each sur-
vey year and the next, the increase was particularly signif-
icant between 1997 and 2000. This may partly be
explained by the favorable prices for maize following a
poor maize crop in the 1998/99 season; as a result, the fer-
tilizer-to-maize price ratio was lower at the start of the
planting season for year 2000, an incentive to farmers to
increase their fertilizer use during that season (using
naive price expectations based on recent output prices).
Figure 16.6 shows the existence of regional differences in
the yield-fertilizer nexus for the period covered by the
household surveys.
Figure 16.6 shows clear differences in the potential for
maize production in the two regions. Even without fertiliza-
tion, yields for unfertilized plots in the high-potential
region are generally higher than those from fertilized plots
in the low- potential region, indicating that the effects of
fertilizer use vary according to other factors.
Table 16.4, which expresses yields in metric tons per acre,
can be used to show why the maize yield estimates reported
by the Ministry of Agriculture are different from the com-
posite yield measure used in this paper. The Ministry of
Agriculture estimates are generally lower than the compos-
ite yield (that is, the sum of the yields for maize and other
crops in the same plot converted to an index using price
ratio of each crop to that of maize), and they do not reflect
the true returns to inputs applied to the plot or field. The
composite yield measure increased from 0.9 ton in 1997 to
1.4 tons in 2007, whereas the standard measure by the Min-
istry of Agriculture, increased from 0.6 to 1.0 ton per acre
over the same years.
Trends in wholesale price margins
Price margins can indicate the state of competition and
innovations that reduce marketing costs between two
points of interest by improving market efficiency. As shown
in figure 16.7, the margin between the wholesale world
price of fertilizer, as measured by the cost, insurance, and
freight (CIF) of fertilizer arriving in the port of Mombasa
on the east coast of Kenya, and prices in the hinterland
town of Nakuru has declined dramatically in the years
between 1990 and 2008.
The world price of fertilizer remained fairly constant from
1990 to 2007, when it rose sharply (see figure 16.7). During
this period the wholesale cost of fertilizer in Nakuru steadily

276 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
Figure 16.5 Trend in Maize Yields in Fertilized and Unfertilized Plots in Kenya for Different Seed Technologies

a. Plots not using fertilizer b. Plots using fertilizer

2,000

1,500
Kilograms per acre

1,000

500

0
97

00

04

07

97

00

04

07

97

00

04

07

97

00

04

07
19

20

20

20

19

20

20

20

19

20

20

20

19

20

20

20
Hybrid seed Nonhybrid Hybrid seed Nonhybrid

Source: Authors, estimated from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007.

Figure 16.6 Trend in Maize Yields for Fertilized and Unfertilized Plots in Kenya, by Region

a. High-potential region b. Low-potential region

2,000

1,500
Kilograms per acre

1,000

500

0
97

00

04

07

97

00

04

07

97

00

04

07

97

00

04

07
19

20

20

20

19

20

20

20

19

20

20

20

19

20

20

20

Unfertilized Fertilized Unfertilized Fertilized

Source: Authors, estimated from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007.

declined, implying that the costs of marketing fertilizer there holders suggest this reduction is a result of increased compe-
had declined, leading to lower prices at Nakuru. Studies tition after reforms in the 1990s, economies of scope result-
(Kimuyu 1994; Wanzala, Jayne, and Staatz 2002; Allgood and ing from mergers, and access to competitive credit from
Kilungo 1996; IFDC 2001), based on interviews with stake- international sources.

CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 277
 Table 16.4 Trend in Composite and “Standard” Maize Yields in Kenya, by Region and Zone
(tons per acre)

1997 2000 2004 2007

High-potential region
Western transitional 0.7(0.5) 1.2(0.8) 1.0(1.0) 1.6(1.2)
High-potential maize zone 1.3(1.1) 1.3(1.0) 1.6(1.4) 1.6(1.3)
Western highlands 0.7(0.5) 1.7(0.9) 0.9(0.8) 1.4(0.9)
Central highlands 1.2(0.6) 1.8(1.1) 1.7(1.0) 2.3(1.3)
Subtotal 1.1(0.8) 1.4(1.0) 1.4(1.1) 1.7(1.2)
Low-potential region
Coastal lowland 0.5(0.2) 1.0(0.6) 0.5(0.3) 0.8(0.5)
Eastern lowland 0.4(0.2) 1.0(0.5) 0.6(0.4) 0.8(0.5)
Western lowland 0.5(0.3) 0.7(0.4) 0.5(0.3) 0.9(0.7)
Subtotal 0.5(0.2) 0.9(0.5) 0.5(0.3) 0.9(0.6)
Grand total 0.9(0.6) 1.3(0.8) 1.2(0.9) 1.4(1.0)

Source: Composite yield estimated by authors from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004,
and 2007. Maize-only yield from Kenya Ministry of Agriculture.

Figure 16.7 Price of Diammonium Phosphate in The analysis conducted for this paper used regression
Mombasa and Nakuru methods including random effects (RE), fixed effects (FE),
and correlated random effects (CRE) to model fertilizer
4,400
Cost per 50 kilogram bag (constant 2007 K sh)

demand (application rate per acre). The RE approach

4,000 assumes strict exogeneity between explanatory variables
3,600 and composite error term, which includes unobserved
household-specific heterogeneity. On the other hand, FE
3,200
does not assume strict exogeneity but takes the unob-
2,800 served effects as constant over time and uses a differencing
2,400 approach to remove these effects to generate consistent
2,000 estimates. Unlike FE, the CRE method extends the RE
analysis by modeling unobserved heterogeneity using the
1,600
household means of time-varying variables. Therefore,
1,200 with CRE, it is possible to test whether the model captures
800 unobserved effects and to use estimates of these effects to
classify households or explain differences between house-
90

92

94

96

98

00

02

04

06

08
19

19

19

19

19

20

20

20

20

20

holds (Wooldridge 2002). An additional benefit of CRE is

Nakuru, wholesale
Source: Kenya Ministry of Agriculture.
HOUSEHOLD AND OTHER DETERMINANTS OF
FERTILIZER DEMAND
Using variables including education, value of assets, land
size, land preparation technology, gender of the household
head, as well as geographic factors such as distance to fertil-
izer seller, agro-ecological conditions, soil types, and market
conditions, it is possible to conduct a regression analysis of
fertilizer demand in the high-potential and low-potential
regions of Kenya. This analysis will provide a measure of
diversity or heterogeneity in demand across the country and
between different households that face varied surroundings,
knowledge that is important in setting appropriate policy
geared to achieving food security for smallholders.
Mombasa, cif
that the estimates on the time-varying variables are the
same as those in the FE estimation, and unlike in the FE
approach, the effect of time-constant factors (such as
gender and location dummies) are estimated as well (not
differenced away as in FE).
Using CRE, it is possible to reject the null hypothesis of
nonexistence of unobserved heterogeneity, implying that
the FE approach is more appropriate than RE, which
assumes exogeneity. However, CRE regression offers the
benefit of producing the same estimates as FE regression for
time-varying variables, while at the same time providing a
way to model heterogeneity so as to explain differences
across households based on skills and other factors that can-
not be observed or for which data cannot be obtained. For
these reasons, the results of only the CRE method are dis-
cussed here, using a double-hurdle approach.

278 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
 Table 16.5, presents the results of the CRE regressions different sign and magnitude in the market participation
for the high- and low-potential regions for fertilizer mar- and demand equations, unlike the Tobit model, which
ket participation and consumption or use decisions. For assumes the same effect and magnitude in both equa-
the double-hurdle model, the same variable can have tions.

Table 16.5 Fertilizer Market Participation and Demand Using Correlated Random Effects to Model Household
Heterogeneity
High-potential region Low-potential region

Market Consumption Market Consumption

Variables participation (kilograms/acre) participation (kilograms/acre)
Dependent variable (units) (0/1) (kilograms/acre) (0/1) (kilograms/acre)
Price for nitrogen (K Sh/kilogram) 0.015 –0.831* –0.009 –0.175**
(0.005) (0.439) (0.009) (0.081)
Price for maize grain (K Sh/kilogram) –0.003 0.313 –0.003 –0.016
(0.003) (0.218) (0.004) (0.026)
Age of household head (years) –0.002** –0.001
(0.001) (0.001)
Quintiles for value of household assets:
2 0.012 0.853 –0.047 0.359*
(0.016) (1.421) (0.029) (0.206)
3 0.001 1.108 0.016 0.571***
(0.019) (1.554) (0.031) (0.219)
4 –0.017 –0.101 –0.001 0.401*
(0.022) (1.704) (0.032) (0.241)
5 –0.012 2.454 0.021 0.897***
(0.024) (1.849) (0.036) (0.261)
Quintiles for total cropped land:
2 0.021 –3.767*** 0.007 0.130
(0.019) (1.394) (0.026) (0.202)
3 0.038* –4.270*** –0.006 –0.326
(0.021) (1.505) (0.027) (0.216)
4 0.068*** –3.995** –0.007 0.058
(0.022) (1.680) (0.028) (0.235)
5 0.069*** –1.594 0.006 –0.261
(0.025) (1.926) (0.032) (0.264)
Categories for education of head of household:
2 1–4 years –0.018 1.475 –0.066 –0.305
(0.022) (2.696) (0.035) (0.338)
3 5–8 years –0.017 0.546 –0.036 –0.155
(0.022) (2.640) (0.034) (0.289)
4 9–12 years 0.001 5.605* 0.043 0.379
(0.024) (2.897) (0.044) (0.357)
5 > 12 years 0.032 6.416* 0.100 0.133
(0.029) (3.698) (0.067) (0.537)
Categories for land preparation technology:
2 Oxen 0.101*** 4.330** –0.016 0.011
(0.021) (1.768) (0.027) (0.200)
3 Tractor 0.147*** 5.670*** –0.030 0.056
(0.020) (1.701) (0.040) (0.293)
Categories for land tenure:
2 Own land without title 0.005 –0.088 0.027 0.068
(0.014) (1.069) (0.018) (0.142)
3 Renting land 0.054*** –1.047 0.030 –0.299
(0.018) (1.508) (0.036) (0.303)
Dummy (1= female head of household) –0.023 –0.647 –0.060*** –0.281*
(0.018) (1.669) (0.020) (0.197)
Categories of soil types:
2 0.007 –1.595 –0.162** –0.479
(0.040) (4.346) (0.079) (0.872)
(continued next page)

CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 279
 Table 16.5 (continued)
High-potential region Low-potential region

Market Consumption Market Consumption

Variables participation (kilograms/acre) participation (kilograms/acre)
3 0.020 –1.643
(0.028) (2.903)
4 0.009 –1.708 –0.008 –0.143
(0.024) (2.577) (0.075) (0.578)
5 –0.094* –7.913* –0.083 –0.377
(0.054) (4.362) (0.103) (0.907)
Agro-zone dummies (central lowland and west transitional dropped):
3 Eastern lowlands –0.036 –0.195
(0.064) (0.710)
4 Western lowlands 0.389*** 2.324**
(0.110) (1.133)
6 High-potential maize 0.398*** –0.435
(0.050) (3.891)
7 West highlands 0.263*** –2.239
(0.059) (4.134)
8 Central highlands 0.423*** 19.053***
(0.058) (6.189)
Dummy (1= single crop in plot) –0.061*** –0.307 –0.001 –0.042
(0.016) (1.084) (0.018) (0.161)
Mundlak–Chamberlain device:
Price for nitrogen (K Sh / kilogram) –0.064** –1.246 –0.062 0.525
(0.025) (2.880) (0.061) (0.576)
Price for maize grain (K Sh/kilogram) 0.053*** –1.669 0.030*** 0.098
(0.011) (1.075) (0.010) (0.092)
Dependency ratio (dependants to productive –0.001 –0.359 0.002 –0.005
members) (0.002) (0.287) (0.004) (0.039)
Distance to fertilizer seller –0.023*** 0.641 –0.009*** –0.012
(0.005) (0.629) (0.003) (0.025)
Duration as head of household (years) 0.004** 0.001
(0.002) (0.002)
Quintiles for value of household assets:
2 0.015 4.074 0.096 –0.495
(0.036) (4.228) (0.065) (0.561)
3 0.088** –3.910 0.046 –0.396
(0.038) (4.261) (0.061) (0.537)
4 0.111*** 0.454 0.058 0.218
(0.040) (4.353) (0.062) (0.567)
5 0.103** –1.483 0.057 –0.064
(0.044) (4.683) (0.064) (0.593)
Quintiles for total cropped land:
2 –0.046 –2.737 0.190*** –1.039
(0.037) (4.147) (0.068) (0.548)
3 –0.002 4.387 0.238*** –0.278
(0.038) (4.129) (0.065) (0.529)
4 –0.067* 0.031 0.173*** –0.489
(0.038) (4.117) (0.064) (0.562)
5 –0.007 2.099 0.178*** –0.657
(0.043) (4.483) (0.068) (0.585)
Fractions of 20-day periods with
< 40 millimeters of rain in season –0.185*** –26.845** 0.100 6.206
(0.124) (12.346) (0.254) (2.479)
Observations (plots) 4,051 4,051 1,782 1,782
Source: Authors.
Note: Quintiles range from lowest (1) to highest (5) for asset values and land sizes (estimates shown in the table are compared with lowest quintile,
which is dropped). Estimates for the remainder of categories (land preparation, zones, tenure, and others) should be interpreted in relation to the
omitted category (category 1). Standard errors are in parentheses; *** p < 0.01, ** p < 0.05, * p <0.1.

280 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
Effect of fertilizer and maize prices on
demand for fertilizer
Results of the regression analysis presented in table 16.5
show that when fertilizer prices increase by approxi-
mately 278 shillings per 50 kilogram bag, household fer-
tilizer application rates (kilograms per acre) decline by
4.4 kilograms (high-potential region) and 1.1 kilograms
(low-potential region). For the high-potential region, this
translates to a response elasticity of 0.52 (that is, a 10 percent
increase in fertilizer price leads to a 5.2 percent decline in
application rates per acre). The elasticity for low-potential
region is 0.32.
From the Mundlak-Chamberlain device (Wooldridge
2002), it is evident that households facing higher fertilizer
prices have lower probability of participation, while those
facing higher maize output prices have higher probability of
participation in fertilizer markets. There is a reduction in
probability of market participation of 0.64 points for
households facing a 10 K Sh-per-kilogram higher fertilizer
price (or K Sh 500 per 50-kilogram bag), while the proba-
bility of participation by households facing a similar higher
maize output price is higher by 0.53 points for the high-
potential region (0.30 in low-potential region)—that is,
households facing higher fertilizer prices have lower proba-
bility of market participation, while those facing higher out-
put prices have higher probability. However, the effect of
prices on the decision about how many kilograms of fertil-
izer to apply per acre is not significantly different across
households. This finding implies that price signals are
important in determining farmers’ decision to participate in
fertilizer markets.
Effect of household resource endowments on
demand for fertilizer
The analysis also takes into account the effect of some mea-
sures of resource endowment (asset values and land size) on
fertilizer demand. Although the probability of market par-
ticipation for the low-potential region does not change with
value of assets, the application rate rises with assets. Raising
a household’s assets from the lowest quintile to the third
quintile or the fifth quintile, for example, raises the applica-
tion rates by 0.571 and 0.897 kilogram per acre, respectively.
But when contrasting different households in the high-
potential region, it is evident that households with assets of
a higher value have a greater probability of participating in
the fertilizer market.
A related measure of resources is the amount of land
under crops during the season (because data used for this
CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
analysis do not include total land owned for the year 2000,
land under crops is used as a proxy for this variable).
Although the probability of market participation rises with
an increase in the size of land for households in the high-
potential region, the amount applied per acre decreases with
land size. Households with more land under crops in the
low-potential region have a higher probability of market
participation, although the rates per acre are not signifi-
cantly different from those with less land under crops.
Effect of land tenure, gender, land
preparation technology, and mixed
cropping on demand for fertilizer
Households that rent land have a higher probability of fer-
tilizer market participation (by 0.05 points) than those that
have title to their land, but differences in application rates
across land tenure categories are not significant. Female-
headed households have a lower probability of market par-
ticipation (0.06) and intensity per acre (0.28) compared
with male-headed households in the low-potential region.
Although signs of estimates are the same as those in the
high-potential region, gender estimates for the high-poten-
tial region are not significant. The authors find no plausible
explanation why the low-potential region has a gender effect
while the high-potential region does not.
In terms of agro-ecological zones, households in the cen-
tral highlands have a 0.4-point higher probability of partic-
ipating in fertilizer markets than those in the western tran-
sitional zone and apply 5.95 kilograms more fertilizer per
acre. On land preparation technology, there is a higher
probability of market participation for households using
animal draught (0.1) or tractors (0.14) compared with
manual hoeing; fertilizer application per acre for these
households also increases by 4.4 and 5.6 kilograms, respec-
tively, over manual technologies. Households that do inter-
crop or mix cropping (that is, plant maize with one or more
other crops in the same plot) have a higher probability of
fertilizer market involvement than those growing only
maize. However, application rates do not significantly differ
with intensity of mixed cropping.
Effect of distance to fertilizer seller, education, and
experience on demand for fertilizer
Households that are further away from fertilizer sellers have
lower probability of participating in the market (0.023 points
per kilometer for the high-potential region and 0.01 points
for low-potential region). In other words, households in the
high-potential region that are located 10 kilometers away
281
 from fertilizer retailers have a 0.23-point lower probability
of participating in fertilizer markets, but application rates per
acre do not differ significantly. Households with head who
have more years of schooling have a positive but insignificant
probability of participating in the market than those who do
not. The intensity of application per acre is positive with years
of education (and significant for high-potential region). Age
of household head in the high-potential region has a negative
relationship with probability of participation but not the
intensity of application. The number of years or duration as
the head of household has a positive effect on probability of
participation in the market in the high-potential region;
experience as head of household raises chances of market
participation.
LESSONS LEARNED, SUSTAINABILITY,
AND POTENTIAL FOR REPLICABILITY
When examining the factors driving growth in fertilizer use
and maize productivity in Kenya from the early 1990s to
2007, the basic story is one of synergies between the liberal-
ization of input and maize markets on one hand and public
investments in support of smallholder agriculture on the
other, a situation that has led to tangible private sector
investment in fertilizer retailing and maize marketing,
which in turn has encouraged an impressive rise in fertilizer
use and maize yields on smallholder farms over the period
1997–2007. This narrative is complicated, however, by the
many changes that Kenya’s economy and business environ-
ment has experienced during this period, both positive and
negative, which have also undoubtedly affected the incen-
tives of farmers, consumers, and private marketing agents.
These factors may not be directly linked to the fertilizer and
maize markets, but their influence on observed indicators
cannot be analytically separated from those of the reforms
highlighted in this paper. However, it is reasonable to
assume that these influences outside the agricultural sector
are of second-order magnitude, compared with the more
direct agricultural policy reforms and investments, in
explaining the behavioral responses of farmers and fertilizer
and maize marketing agents.
There are several pathways through which government
actions in fertilizer and maize markets has positively
affected the agricultural sector and rural and urban living
standards in Kenya in recent years. As shown in figure 16.2,
the government of Kenya implemented a number of policy
reforms affecting the incentives for investment by private
fertilizer distribution firms. The government also legalized
domestic and regional maize trade, although other actions
282 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
during the 1990s partially eroded the potential response by
the private sector. Despite the mixed government stance
toward maize market liberalization during the 1990s and
early 2000s, evidence of increased private sector investment
is tangible. Traders buying maize directly from farmers have
penetrated more deeply into smallholder areas. Increased
competition and efficiency in maize milling and retailing is
also evident in the significant decline in maize marketing
margins. There is also strong evidence of increased state
investment in public goods supportive of private sector
investment, especially since the creation of the Constituency
Development Fund (CDF) in 2003. The combination of
supportive policy changes in the fertilizer, foreign
exchange, and maize markets, coupled with improved
access to markets and services made possible by public
good investments, appears to have stimulated investment
by the private sector in both maize and fertilizer marketing.
These factors have worked synergistically to bring about
important gains in maize productivity and benefits to
smallholder farmers and consumers in Kenya.
Evidence of increased smallholder fertilizer use and maize
yields is drawn from nationwide household panel data from
four surveys conducted by Egerton University’s Tegemeo
Institute between 1997 and 2007. Because the data constitute
a balanced nationwide panel of 1,260 households,5 the
results provide a fairly reliable indicator of the changes in
fertilizer use patterns over time, although the surveys are not
strictly nationally representative. The main findings of the
surveys are as follows:
■ The percentage of sampled smallholders using fertilizer
on maize increased from 56 percent in 1996 to 70 percent
in 2007.
■ Fertilizer application rates (for all maize fields including
unfertilized fields) rose from 34 kilograms an acre in
1997 to 45 kilograms an acre in 2007, a 32 percent
increase.
■ There are wide regional variations in fertilizer use. More
than 90 percent of smallholder farmers use fertilizer on
maize in three of the broad zones surveyed: the high-
potential maize zone, western highlands, and central
highlands. Fertilizer use is low and barely rising in most
of the semiarid regions (coastal and western lowlands,
and the marginal rain shadow). However, fertilizer use
has risen impressively in the medium-potential eastern
lowlands and Western Transitional zones, where the
percentage of households using fertilizer on maize rose
from 21 and 39 percent, respectively, in 1997 to 43 and
81 percent in 2007.
■ While the total area under maize in Kenya remained
largely constant over 1997–2007, maize yields increased
by roughly 18 percent during the same period. This yield
improvement is not borne out in official government
maize production statistics, however, which do not take
into account the shift over time in the proportion of
maize area grown under intercropped cultivation or the
shift over time in the proportion of maize area grown in
relatively semiarid regions, which has been facilitated by
the release of improved maize cultivars well suited to
mid- and low-altitude areas of the country. To assess
changes in maize yield, it is important to account for the
gradual shift in the proportion of maize area under
monocropped versus intercropped cultivation as well as
the expansion of maize production in the more semiarid
parts of the country. After stratifying between hybrid and
nonhybrid users and between intercropped and
monocropped maize fields, the household survey data
show that maize yields on all types of fields have
increased over time, reflecting the influence of many fac-
tors in addition to fertilizer use. Fertilizer use and maize
yields have increased especially rapidly on the inter-
cropped fields, less so on monocropped fields.
■ Fertilizer marketing costs declined substantially in con-
stant K Sh between the mid-1990s and 2007. Interviews
with key informants in Kenya’s fertilizer sector identified
four factors responsible for the declining fertilizer mar-
keting costs observed in Kenya: the potential for cheaper
backhaul transportation has been exploited by making
more use of trucks transporting cargo from Rwanda and
Democratic Republic of Congo to the port of Mombasa;
private importers are increasingly using international
connections to obtain credit at lower interest rates and
financing costs than are available in the domestic econ-
omy; local and international firms have merged, enabling
shared knowledge and economies of scope that save local
distribution costs; and increased competition among
local importers and wholesalers has expanded the num-
ber of firms engaged in fertilizer marketing since the
early 1990s. It is likely that the fourth factor—increased
competition—has to some extent stimulated firms to
exploit the other cost-reducing innovations identified in
order to maintain their market position.
To assess the robustness of the Tegemeo Institute’s sur-
vey findings, the proportion of smallholder households
purchasing fertilizer according to the survey results was
compared with estimates based on three other analyses
covering a subset of the same districts during the same
CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
general time period. The Tegemeo survey estimates are
comparable and in some cases lower than other estimates of
fertilizer purchases and dose rates. The rise in smallholder
use of fertilizer in the Tegemeo survey data is also consis-
tent with official Ministry of Agriculture figures (shown in
figure 16.1), which indicate that total fertilizer consump-
tion in Kenya rose 65 percent between 1997 and 2007.
The rise in fertilizer use in Kenya has not been uniform
across regions, however. Use rates are much higher in areas
where the main-season rainfall is relatively high and stable
than they are in the drier areas. Fertilizer use is highly risky
in many of the semiarid regions, where environmental fac-
tors are likely to limit the role of fertilizer in contributing to
poverty alleviation and food security unless it is accompa-
nied by actions to improve soil organic matter and moisture
(Marenya and Barrett 2008). Within a given agro-ecological
zone, it is evident that the decision of households to pur-
chase fertilizer is only slightly related to farm size and unre-
lated to household wealth. In relatively productive areas, the
proportion of poorer and wealthier households applying
fertilizer on maize is similar. In risky environments, only a
small proportion of either poor or wealthy households use
fertilizer on maize.
These gains in smallholder fertilizer use and maize yields
have been encouraged by Kenya’s decision to liberalize input
and maize markets in the early 1990s. New entries and
investment in fertilizer wholesaling and retailing have been
massive since the early 1990s. The International Fertilizer
Development Center (IFDC) estimates that more than
500 wholesalers and 7,000 retailers are operating in the
country. This has led to a denser network of rural retailers
and a major reduction in the distance between farms and
fertilizer sellers, which has contributed to the impressive
growth in fertilizer use by Kenyan smallholders from the
early 1990s to 2007. The Tegemeo Institute survey data also
indicate that the mean distance traveled by farmers to sell
their maize to private traders declined over 1997–2007; the
median distance as of 2007 was zero, indicating that assem-
bly traders tend to purchase maize right from farmers’
fields. Analysis of wholesale maize grain prices and retail
maize meal prices indicate that the miller-retail marketing
margin has declined significantly over time, conferring
benefits mainly to consumers. More than 50 percent of
rural farm households are either buyers or net buyers of
maize, while virtually all urban households purchase
maize meal each year (Mukumbu and Jayne 1994; Jayne
and Argwings-Kodhek 1997).
Other signs of improvement in maize markets in Kenya
include farmers’ level of satisfaction with the performance
283
 of maize markets from their subjective perspective. More
than 65 percent of farmers surveyed in the nationwide Tege-
meo Institute rural surveys indicated that they prefer the
current liberalized maize marketing system to the former
controlled marketing system, primarily because grain is eas-
ier to sell, farmers are paid in cash at the time of sale, and
maize is more reliably available for purchase.
In 2008, however, the positive trends in Kenya’s maize
and fertilizer markets were reversed by civil disruption,
drought, and the unprecedented surge in world fertilizer
prices. The civil unrest led to the destruction of much phys-
ical infrastructure in western Kenya, such as petrol stations
and grain storage, as well as to the closing of many input
supply stores, in early 2008. Moreover, incentives to use fer-
tilizer in Kenya have been adversely affected both by
drought and world events as maize/fertilizer price ratios
plunged to their lowest level in at least 18 years. Figure 16.8
plots monthly wholesale maize to wholesale fertilizer price
ratios per ton in Nakuru. The higher the ratio, the more
profitable the maize, and therefore the greater the incentive
to apply fertilizer on maize. While this ratio has historically
ranged between 0.4 and 0.6 at the time of planting, in 2008
it plunged to below 0.25 because of the increase in world
fertilizer prices. The price of maize in Kenya has not risen
nearly as dramatically as fertilizer.
These findings have implications for policy options.
The main general lesson is the need for a public-private
Figure 16.8 Maize/Fertilizer Price Ratios, Nakuru, Kenya,
1994–2008
0.8
tonne maize/tonne DAP fertilizer price ratio
relationship that encourages investment in input and out-
put marketing services for smallholder farmers. In Kenya’s
case, this relationship was achieved through a combination
of investments in public goods and institutional reforms
supportive of liberalized marketing, even though the
maize marketing reforms were at times subject to rever-
sals. Considerable additional gains could be made in
smallholder and consumer welfare if progress could also
be made in the following areas:
Consider changes in government actions in the
transport sector that could reduce fertilizer and
grain distribution costs. For example, because of fre-
quent delays in off-loading of commodities at the port of
Mombasa and because of the erosion of the regional railway
system, it is difficult to arrange for upcountry transport of a
full shipload of fertilizer. Because of this coordination prob-
lem, fertilizer importers have invested in storage facilities
near the port where fertilizer can be temporarily stored
until trucks arrive for loading and upcountry distribution.
These investments make sense if upland transport con-
straints and the delays and inefficiency at the Port of
Mombasa are taken as given. However, if procedures for
streamlining the efficiency of off-loading at the port could
be achieved (for example, by privatizing stevedore services
and issuing performance contracts or by devolving wider
management of port operations to professional firms), thus
reducing off-loading time and the storage costs incurred at
Mombasa for lack of sufficient transport, then fertilizer
importing firms could avoid these extra charges. In a com-
petitive marketing environment, these reductions in fertil-
izer marketing costs would then be passed along in the form

of lower farm gate prices.6

0.7

Reduce transaction costs associated with VAT and

0.6
port operations. Currently fertilizer, as well as most other
0.5 farm inputs, is zero-rated with respect to import duties.
This means that no duty is charged on fertilizers, although
0.4 at least until 2007, a VAT on related services was still levied.
A VAT is charged, for example, on transport and services
0.3
such as bagging at the port of Mombasa. Although the VAT
0.2
is supposed to be refunded, the process is lengthy and is a
source of continuing frustration for market participants. In
0.1 addition, port handling charges, Kenya Bureau of Stan-
dards charges, and other taxes account for 17 percent of
94

96

98

00

02

04

06

08
19

19

19

20

20

20

20

20

CIF (Gitonga 2004). Port fees, levies, and accessorial charges

Source: Kenya Ministry of Agriculture, Market Information Bureau. need to be rationalized and aggregated. In addition, the
Note: Price ratio is defined as the wholesale market price per ton of maize
in Nakuru divided by the cost of diammonium phosphate (CIF) in Nakuru numerous documentation procedures need to be reduced
per ton, in nominal Kenyan shillings. and, if possible, provided through electronic means.

284 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
Interviews with key informants in the fertilizer industry
have identified numerous other potential sources of cost
savings, many of which require action on the part of gov-
ernment to improve efficiency.
Invest in rehabilitating the eroded rail, road, and
port infrastructure, which would reduce distribu-
tion costs. The farm gate price of fertilizer in western
Kenya is roughly twice as high as the landed cost at Mom-
basa, and transport costs are the major component of this
cost difference. High farm gate prices of fertilizer restrict
demand for its use and depress agricultural productivity.
Hence, efforts to improve the efficiency of port costs and
upland shipping would bring major economywide benefits.
In particular, rail transport could reduce these costs sub-
stantially and also save government spending on repairing
roads damaged by heavy truck traffic.
Tailor fertilizer packages to local demand condi-
tions. This action would increase demand from smaller
farmers who require and are able to purchase only small
packets. Repackaging of fertilizers from 50 kilogram packets
into packets of 25, 10, 2, and 1 kilograms is increasingly
taking place, but the process is sometimes associated with
fertilizer adulteration and counterfeit products. (That
said, adulteration and sales of counterfeit products are
often isolated events rather than a well-organized activi-
ties, according to Global Development Solutions.)7 Part of
the wide fluctuation in the nitrogen and phosphorous
concentration in fertilizers can be attributed to the
absence of effective measurement and calibration facilities.
In this context, the Kenya Plant Health Inspectorate Ser-
vice and the Kenya Pesticide Board should become more
effective in monitoring and controlling adulteration and
counterfeit products, as well as in intensifying farmer and
retailer awareness programs to help protect farmers from
substandard products.
Raise fertilizer response rates through agronomic
training of farmers. The profitability of fertilizer use
could be enhanced by improving the aggregate crop yield
response rates to fertilizer application. This requires mak-
ing complementary investments in training for farmers
on agronomic practices, soil fertility, water management,
and efficient use of fertilizer and investing in crop science
to generate more fertilizer-responsive seeds.8 Emerging
problems of soil acidity in the maize belt of western Kenya
indicate that soil pH levels may need to be raised to
ensure profitable use of fertilizer in these areas. Survey
CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
data commonly indicate that the contribution of fertilizer
to food grain yields varies tremendously across farms
even within the same villages. Simply bringing fertilizer
response rates among the bottom half of the distribution
up to the mean would result in substantial improvements
in household and national food security (Nyoro, Kirimi,
and Jayne 2004).
Producer organizations, despite their poor track
record, will be increasingly important for rural
income growth. Assuming that the management prob-
lems and politicization of producer organizations and
cooperatives could be minimized, they might afford an
important pathway for smallholders to achieve higher levels
of input use and to adopt better production and marketing
practices than the current separate and uncoordinated
stages in the supply value chains. The role of independent
producer groups would be to reduce the transaction costs
and risks of private marketing firms dealing with farmers
and to develop a production base through the transfer of
credit, inputs, and know-how. The Farm Inputs Promotions
and the Kenya Market Development Program/Cereal Grow-
ers Association farmer training programs are examples of
successful attempts by the government, development part-
ners, and NGOs to assist and train groups and to utilize
farm extension knowledge, supply chain development, and
fertilizer technologies.
While all of these measures can contribute to increased
fertilizer use, none is likely to prove effective on its own. Pol-
icy makers should, therefore, select strategic combinations
of supply- and demand-side measures to allow supply and
demand to grow in parallel—strengthening the basis for
viable private sector-led commercial fertilizer markets.
The final question is about the role of fertilizer subsidies.
The greatest scope for subsidies to promote fertilizer use is
in areas where fertilizer use is far below its optimal levels
after taking into account the maize yield response to fertil-
izer and the riskiness of applying fertilizer, especially in
semiarid regions where crop failure is not unusual. Recent
evidence indicates that crop response to fertilizer applica-
tion varies widely among smallholder farmers even within
the same villages because of differences in management
practices, soil quality, timeliness of application, and so
forth. The evidence also shows substantial scope for raising
the efficiency of fertilizer use, at least for farmers who are
currently getting lower response rates from fertilizer appli-
cation than their more efficient neighbors (Marenya and
Barrett 2009; Xu et al. 2009). Moreover, there is little empir-
ical evidence to determine how prevailing levels of fertilizer
285
 application compare with optimal levels when taking these
factors into account. Fertilizer use rates are clearly low in the
semiarid areas of Kenya, and fertilizer subsidies in these
areas would likely raise fertilizer use, but the contribution to
yields and smallholder incomes may be quite limited
because of the environmental riskiness and low response
rates in such areas. A major question for semiarid areas,
therefore, is whether poverty reduction and food security
objectives can be best achieved through fertilizer subsidies
or other types of public programs and investments. Given
that resources are scarce, efforts should be made to identify
the types of agricultural expenditures that will generate the
greatest payoffs.
In the high-potential areas, a large majority of farmers is
already purchasing fertilizer. Although use rates were quite
high in 2007, they are likely to have fallen given the ensuing
adverse conditions since then. Fertilizer subsidies are polit-
ically attractive in that they promise increased fertilizer use
and food production, but these outcomes are by no means
assured. In 2009 Kenya faced its lowest maize production
level in recent history after having initiated a major fertil-
izer subsidy program; poor rains in 2009 rendered the fer-
tilizer subsidy program relatively ineffective, leading the
country to import more than 1 million tons of maize in
2009. Moreover, providing subsidized fertilizer in areas of
high commercial demand will almost certainly result in a
partial crowding out of commercial sales, as shown by the
findings of studies conducted in Zambia and Malawi,
where commercial demand for fertilizer is considerably
lower than in Kenya (see Xu et al 2009; Dorward et al.
2008). Where purchase of commercial fertilizer is high,
then a ton of subsidized fertilizer distributed by govern-
ment is unlikely to result in an additional ton of fertilizer
being applied on farmers’ fields, because the farmers previ-
ously purchasing fertilizer are no longer likely to buy it if
they can acquire the same amount more cheaply from a
government program.
In the current high price environment, the availability
of seasonal loans for input purchases takes on heightened
importance for maintaining farmers’ effective commercial
demand for fertilizer. Many Kenyan farmers have been
able to finance fertilizer through the credit offered in the
integrated input-output chains for crops such as tea,
sugar, and coffee. These integrated marketing arrange-
ments have also provided the means for farmers to obtain
fertilizer for their food crops, since the companies can
recoup their loans for other crops as well when the farm-
ers sell their cash crop back to the company. But in areas
where fertilizer use on a particular crop is profitable, such
286 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
as maize in western Kenya and horticulture throughout
the country, most farmers have achieved reasonable levels
of fertilizer use without credit. Support for the develop-
ment of viable credit programs may also help smallholder
farmers maintain their access to fertilizer use despite cur-
rent high prices for households in which liquidity con-
straints are the main problem.
The experience of Kenya demonstrates the role of a sup-
portive policy environment that attracts local and foreign
direct investment in improving smallholder farmers’ access
to input and commodity markets. In Kenya’s case, a stable
input marketing policy environment has fostered a private
sector response that supports smallholder agricultural pro-
ductivity and poverty alleviation. These goals remain elusive
in countries lacking a sustained commitment to the devel-
opment of viable commercial input delivery systems. While
the government’s policy stance toward maize marketing has
been prone to vacillation, the operations of the NCPB and
the elimination of regional trade barriers since the inception
of the East Africa Community Custom Union in January
2005 have both promoted maize price stability (Jayne,
Myers, and Nyoro 2008; Chapoto and Jayne 2009). Comple-
mentary programs to support small farmer productivity,
such as the Farm Inputs Promotion program, the agro-
dealer training and credit program, and the organization
of farmers into groups to facilitate their access to exten-
sion and credit services under the Kenya Market Develop-
ment Program have also been important factors in raising
fertilizer use in Kenya.
Because mean household incomes are higher and infra-
structure relatively better in Kenya than in many other
African countries, the market-led growth in smallholder
fertilizer use in Kenya may not be easily transferable to
countries where effective demand is highly constrained.
Kenya’s success in increasing fertilizer usage among small-
holder farmers is also tenuous. Sustaining the momentum
will depend on continued public investment, good policy
choices, favorable weather conditions, and avoidance of
international events detrimental to Kenya. Governance
problems and civil disruption are jeopardizing the sustain-
ability of the commercially driven input distribution system
and rural development more generally. Continued access to
input credit for small farmers in many parts of the country
will require government commitment to limit the potential
for politicization and interference in the management of the
interlinked crop marketing systems for sugarcane, tea, and
coffee, which have provided a means for farmers to acquire
additional fertilizer on credit for use on food crops. Also,
new investment is needed in Kenya’s eroded rail, road, and
port infrastructure if Kenya is to maintain its competitive-
ness. Last, effective systems to improve smallholders’ crop
husbandry and management practices are needed to provide
incentives for continued expansion of fertilizer use and pro-
ductivity growth in areas where fertilizer is only marginally
profitable at present.
NOTES
1. “Kenya to Import 1.5 Million Bags of Cheap Fertilizer,”
Daily Nation, June 23, 2009, p. 33. Throughout this chapter,
“tons” are metric tons.
2. The high-potential region has higher productive poten-
tial and covers the agro-zones of Western and Central High-
lands, High Maize Potential, and Western transition areas,
which include the districts of Bomet, Bungoma, Gishu,
Kakamega, Kisii, Meru, Muranga, Narok, Nakuru, Nyeri,
Trans Nzoia, Uasin, and Vihiga. The low-potential region
consists of the lowland zones in the coast, east, and west of
the country (which are generally drier and have poorer-
quality soil). The low-potential region includes the districts
of Kilifi, Kisumu, Kitui, Kwale, Machakos, Makueni,
Mwingi, Siaya, Taita, and Taveta.
3. The sum of the product of plot fertilizer application
rates and ratio of plot size (acres) to total acres is calculated
for all plots in the household. This procedure gives more
weight to application rates in bigger plots indetermining
aggregate household application rates. Rate = Σ (plot rate *
plot area / total household area).
4. Note that “hybrid” stands for purchased hybrid seed and
open pollinated varieties and “nonhybrid” consists of recy-
cled or replanted hybrids and some “traditional” seed types
of unknown source.
5. In other Tegemeo papers, the balanced panel consists of
1,275 households, but 15 households did not have complete
information on all variables used in this study, hence the
1,260 sample size.
6. Some efficiency improvements in Mombasa port opera-
tions have been recently implemented, and more compre-
hensive reforms are currently under consideration.
7. According to Global Development Solutions (2005),
nearly 3–5 percent of repackaged fertilizers are sold using
counterfeit labels and packages. Specifically, fake brand
name labels are used to sell inferior quality fertilizers.
8. Research indicates that the highest crop yield response is
obtained when improved seed, fertilizer, and agronomic
practices to raise soil organic matter are combined
(Marenya and Barrett 2008; Kelly 2006). In some areas,
improved management practices may have greater impact
on yields than use of fertilizer alone (Haggblade, Tembo,
and Donovan 2004).
CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
REFERENCES
Allgood, J. H., and J. Kilungo. 1996. “An Appraisal of the
Fertilizer Market in Kenya and Recommendations for
Improving Fertilizer Use Practices by Smallholder
Farmers: A Field Report.” International Fertilizer
Development Center (IFDC), Muscle Shoals, AL.
Ariga, J. M., and T. S. Jayne. 2008. “Maize Trade and Mar-
keting Policy Interventions in Kenya.” Paper presented at
the Conference on Food Marketing and Trade Policies in
Eastern and Southern Africa. EST Division, Food and
Agriculture Organization (FAO), Rome, March 1.
Ariga, J. M., T. S. Jayne, and J. Nyoro. 2006. “Factors Driving
the Growth in Fertilizer Consumption in Kenya,
1990–2005: Sustaining the Momentum in Kenya and Les-
sons for Broader Replicability in Sub-Saharan Africa.”
Working Paper 24, Tegemeo Institute, Egerton University,
Nairobi.
Argwings-Kodhek, G. 1996. “The Impacts of Fertilizer
Market Liberalization in Kenya.”
Policy Analysis Matrix Project of the Kenya Marketing
Development Programme. Nairobi.
———. 2004. “Kenya Agriculture Sector Brief.” FAO, Rome.
Chapoto, A., and T. S. Jayne. 2009. Open versus Closed Maize
Border Policy: A Comparison of Maize Price Instability in
East and Southern Africa. International Development
Working Paper. Department of Agricultural, Food, and
Resource Economics, Michigan State University, East
Lansing, MI.
Dorward, A., E. Chirwa, V. Kelly, T. Jayne, R. Slater, and D.
Boughton. 2008. “Evaluation of the 2006/7 Agricultural
Input Supply Programme, Malawi.” Final Report of the
School of Oriental and African Studies, Wadonda Consult,
Michigan State University, and Overseas Development
Institute, undertaken for the Ministry of Agriculture and
Food Security, Government of Malawi, Lilongwe.
Global Development Solutions. 2005. “From Laboratory to
the Dining Table: Tracing the Value Chain of Kenyan
Maize.” Report prepared for the World Bank. Reston, VA.
Gitonga. K. T. 2004. “Study on Rationalization and Harmo-
nization of Policies, Regulations, Procedures, Grades,
and Standards in the Fertilizer Sub-sector in Eastern
Africa.” Kenya Report. Kampala: Eastern and Central
Africa Program for Agricultural Policy Analysis.
Haggblade, S., G. Tembo, and C. Donovan. 2004. “House-
hold Level Financial Incentives to Adoption of Conser-
vation Agricultural Technologies in Africa.” Working
Paper 9. Food Security Research Project, Lusaka.
IFDC (International Center for Soil Fertility and Agricul-
tural Development ). 2001. “An Assessment of Fertilizer
Prices in Kenya and Uganda: Domestic Prices vis-à-vis
International Market Prices.” IFDC, Muscle Shoals, AL.
287
 Jayne, T. S., and G. Argwings-Kodhek. 1997. “Consumer
Response to Maize Market Liberalization.” Food Policy
22 (5): 447–58.
Jayne, T. S., R. J. Myers, and J. Nyoro. 2008. “The Effects of
Government Maize Marketing Policies on Maize Prices
in Kenya.” Agricultural Economics 38 (3): 313–25.
Kelly, V. 2006. “Factors Affecting Demand for Fertilizer in
Sub-Saharan Africa.” Agriculture and Rural Development
Discussion Paper 23, World Bank, Washington, DC.
Kimuyu, P. 1994. “Evaluation of the USAID/Kenya Fertilizer
Pricing and Marketing Reform Program.” U.S. Agency
for International Development, Nairobi.
Marenya, P. P., and C. B. Barrett. 2008. Soil Quality and Fer-
tilizer Use Rates among Smallholder Farmers in Western
Kenya. Ithaca, N.Y: Cornell University.
Ministry of Agriculture, Kenya. 2004. “Strategy for Revital-
izing Agriculture: 2004–2014.” Nairobi.
———. 2008. “Economic Review of Agriculture: 2008.”
Central Planning and Project Monitoring Unit, Nairobi.
Mukumbu, M., and T. S. Jayne. 1994. “Urban Maize Meal
Consumption Patterns: Strategies for Improving Food
Access for Vulnerable Urban Households in Kenya.”
Paper presented at the Symposium on Agricultural
Policies and Food Security in Eastern Africa, Egerton
University, Tegemeo Institute, Nairobi, May 19–20.
288 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS
Nyoro, J., M., W. Kiiru, and T. S. Jayne. 1999. “Evolution of
Kenya’s Maize Marketing Systems in the Post Liberaliza-
tion Era.” Working Paper 2, Tegemeo Institute, Egerton
University, Nairobi.
Nyoro, J., L. Kirimi, and T. Jayne. 2004. “Competitiveness of
Kenya and Ugandan Maize Production: Challenges for
the Future.” Working Paper 10, Tegemeo Institute,
Egerton University, Nairobi.
Tegemeo Institute. 2007. “Enhancing Market Access and
Technology Adoption among Smallholders: A Baseline
Report for Evaluating Agricultural Interventions Funded
by the Rockefeller Foundation in Western Kenya.” Report
prepared for the Rockefeller Foundation, Egerton Uni-
versity, Nairobi.
Wanzala, M., T. S. Jayne, and J. Staatz. 2002. “Fertilizer
Markets and Agricultural Production Incentives:
Insights from Kenya.” Working Paper 4, Tegemeo Insti-
tute, Egerton University, Nairobi.
Wooldridge, J. 2002. Econometric Analysis of Cross Section
and Panel Data. Cambridge, MA: MIT Press.
Xu, Z., B. Burke, T. S. Jayne, and J. Govereh. 2009. “Do Input
Subsidy Programs “Crowd In” or “Crowd Out” Commer-
cial Market Development? Modeling Fertilizer Demand
in a Two-Channel Marketing System.” Agricultural Eco-
nomics 40 (1): 79–94.
 CHAPTER 17
Malawi’s Agricultural Input Subsidy Program
Experience over 2005–09
Andrew Dorward, Ephraim Chirwa, and T. S. Jayne
he implementation in Malawi of a large-scale agri-
T
cultural input subsidy program in the 2005/06
agricultural season and in subsequent years has
attracted significant international interest. 1 While much of
this attention has applauded reported growth in maize pro-
duction and food security in the country, there have also
been significant criticisms and questions. These have
focused on the effectiveness and efficiency of the program
in raising maize productivity, its impacts on the develop-
ment of sustainable commercial input markets (Ricker-
Gilbert, Jayne, and Chirwa 2011), its high and (from
2005/06 to 2008/09) dramatically rising fiscal and macro-
economic costs, its opportunity costs (in terms of crowding
out of other investments), its overall return on investment,
and its sustainability (SOAS et al. 2008; Dorward and
Chirwa 2009a, 2009b, 2010; Dorward, Chirwa, and Slater
2010a, 2010b; Kelly, Boughton, and Lenski 2010).
The importance of agriculture—specifically, maize—to
the Malawian economy and to the livelihoods of most
Malawian people is the critical backdrop to the agricultural
input subsidy program (AISP), together with low agricul-
tural and maize productivity and associated high national,
individual, and household food insecurity.2 Large numbers
of very poor people in Malawi work on very small areas of
Andrew Dorward is Professor of Development Economics at the School of Oriental and African Studies, University of London; Ephraim Chirwa is
Professor of Economics at Chancellor College, University of Malawi; and T. S. Jayne is Professor of International Development at Michigan State
University.
land that are predominantly planted to maize (see table 17.1
for some key indicators). Continual cultivation of maize on
the same land without addition of organic or inorganic fer-
tilizers leads to low yields. Low yields then lead to inability
to afford the purchase of inputs. Purchase of inputs on
credit is also not possible for most farmers because rural
credit markets are underdeveloped and the costs of credit
administration are too high, as are risks for both borrowers
and lenders. Low volumes of input demand, poor infra-
structure, and high transport costs lead to high input costs
and inhibit the development of input supply systems in less
accessible areas, while highly variable maize prices (dis-
cussed below) add to the risks of input use (whether pur-
chased with cash or credit).
Increasing maize productivity is difficult for several rea-
sons. Only 10 percent of Malawian maize producers are net
sellers of maize, while 60 percent are net buyers of maize
(SOAS et al. 2008), and hence most (particularly poorer)
people’s livelihoods and food security are damaged by high
maize prices. Increased maize productivity from the use of
purchased inputs requires, however, that the use of the
inputs is profitable for farmers, and that requires sufficiently
high maize prices and yield responses to cover the costs of
inputs. Unless substantial improvements can be made in
289
 Table 17.1 Key Data on Smallholder Agriculture in Malawi, 2004/05
Indicator North Center South National
Rural population (percent of total population) 10 38 40 88
Income and poverty
Median expenditure/capita (MK thousands) 17 20.9 16.9 17.5
Poor households (percent of rural population) 56 47 64 52
Nutrition and food security
Mean rural daily per capita consumption (kilocalories) for people living
below the poverty line 1,738 1,811 1,703 1,746
Incidence of stunting in children 6 months–5 years (percent) 39.6 47.9 40.8 43.7
Incidence of underweight children 6 months–5 years (percent) 16.1 20 17.2 18.3
Share of calories from own production 0.53 0.58 0.47 0.52
Median month after 2004/05 harvest own food crop exhausted (actual)* — — — 4
Suffered large rise in food prices in past 5 years (percent) — — — 79.2
Smallholder agriculture
Landholding:
Less than 0.5 hectare/household (percent ) 12.1 15.4 25.4 19.9
Less than 1.0 hectare/household (percent ) 31.4 40.6 54.1 46.2
Suffered crop yield loss in past 5 years (percent) — — — 68.8
Maize growers (percent) 93 97 99 97
Access to credit for food crop inputs (percent ) 2.5 4.2 3.0 3.4
Percentage of smallholder farmers purchasing fertilizer (percent) 37 44 39 43
Fertilizer applied on all fields (kilograms)a 32 45 24 34
Fertilizer applied on fertilized maize fields (kilograms/hectare) 139 111 77 101

Source: SOAS et al. (2008) using data from NSO (2005) except * (authors’ calculations from NSO 2006).
Note: — = not available. $1 = MK 140 for most of the years covered in this paper.
a. Fertilizer rates on tobacco plots are roughly double rates across all plots.

yield responses, there is a significant dilemma between the
need for low maize prices for a large numbers of poor maize
buyers (who are also significant maize producers) and the
need for higher maize prices to allow increased returns from
input use to reliably cover the purchase costs. Further diffi-
culties arise from high maize price variability, which dam-
ages both producers and consumers—low prices present
risks to investments in inputs by producers who aim to have
a marketable surplus, while high prices present risks to con-
sumers (including the majority of smallholder farmers).
Poor access to international and domestic markets (caused
in large part by historically low public investment in trans-
port infrastructure), seasonal scarcities, and poor local
market development (resulting from low and uncertain vol-
umes, high costs of transport, and uncertain government
intervention) have led in the past to high intra- and inter-
seasonal maize price variation (as well as higher farm gate
input prices and lower farm gate produce prices), further
depressing market development. Risks of high maize prices
encourage poor consumers to grow as much of their own
staple food as possible, even at very low levels of productiv-
ity. At the same time, there are limited higher-earning
opportunities within or outside agriculture. The result is a
lock-in to low-productivity maize cultivation.
Productivity and investment in productive activities are
further constrained by poverty and by vulnerability to a
wide variety of (often related) shocks, particularly low crop
yields, sickness affecting household members, high food
prices, and losses of employment or remittance income.
Women, who play a key role in agricultural production and
rural livelihoods, tend to be particularly vulnerable to these
shocks. Macroeconomic conditions before 2005—namely,
high real interest rates, high inflation, and significant deval-
uation of the kwacha (MK)—also inhibited growth. How-
ever, macroeconomic management has improved dramati-
cally since 2004.
Agricultural, rural, and national economic development
in Malawi are therefore constrained by a number of inter-
acting household, local, and national vulnerability, poverty
and productivity traps, as illustrated in figure 17.1. These
traps constrain input and maize market development,
investments in maize intensification, diversification out of
maize into other agricultural and nonagricultural activities,
the ability of rural people (particularly the poor) to protect
themselves from shocks, and wider local and national eco-
nomic development. The result is a vicious circle of unsta-
ble maize prices inhibiting net producers’ investment in
maize production, net consumers’ reliance on the market

290 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Figure 17.1 Vicious Circle of the Low Maize Productivity Trap
Unstable
policies
Unstable
weather
Unstable
maize prices
Low producer
investment
Low maize
and
agriculture
productivity
Low and vulnerable
real incomes
Source: Authors.
Note: Light green arrows represent feedback effect.
for maize purchases, and poor consumers’ escape from low-
productivity maize cultivation. These in turn inhibit the
growth of the nonfarm economy. Sustained improvements
in maize productivity with low and stable prices are
required to drive diversification out of low productivity
maize into a more diversified and productive economy that
benefits all Malawians, particularly those who are currently
poor and food insecure.3
Input subsidy and maize market intervention policies
have been a long-standing, major, though often contentious,
feature of strategies of both the government of Malawi and
varied donors’ to promote agriculture and food security.
From the mid-1970s to the early 1990s, the government
financed a universal fertilizer subsidy, subsidized small-
holder credit, and controlled maize prices. The system
began to break down in the late 1980s and early 1990s, how-
ever, with cash flow difficulties, rising treasury deficits, par-
tial market liberalization and increasing importance of
parallel grain markets. The state system of subsidized input
loans, with loan recovery through farmers’ delivery of grain
to the Agricultural Development and Marketing Corpora-
tion (ADMARC), collapsed in the mid-1990s as a result of
the coincidence of widespread harvest failure, multiparty
elections, credit default, the rise of parallel grain markets,
partial implementation of liberalization and structural
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Slow Poor
private roads
sector
development
Consumer
Lock-in
to low-productivity
maize
Low demand for
nonagriculture
goods and services
adjustment policies, and substantial devaluation (raising
local fertilizer prices). With other policy changes drawing
more productive smallholders away from surplus maize
production and into tobacco production, there was a wide-
spread perception in Malawi in the mid-1990s that falling
fertilizer support was leading to diminished maize produc-
tion and a food and political crisis. From 1998/99 the gov-
ernment, with mixed donor support, reinstated a variety of
interventions subsidizing maize fertilizer and seed access,
with intermittent interventions in maize markets. Seed and
fertilizer subsidies shifted from universal price subsidies to
free provision of small “starter packs” initially provided to
all households in 1998/99 and 1999/2000, and then to a
more limited (but varying) number of targeted households
in 2001/02 to 2004/05 (see, for example, Harrigan 2003).
Analysis of smallholder agricultural performance since
the late 1990s is complicated by difficulties with data and in
separating out the effects of poor rainfall and of policy
changes responding to perceptions of an impending food
crisis. In contrast to the widespread perception that maize
production fell during the 1990s and 2000s, official maize
production and overall food production estimates show a
strong rising trend through the 1990s to 2006, together with
a modest rising trend during the same period in per capita
food production (including sometimes disputed estimates
291
 of increasing cassava production). There were two years of
very poor rainfall in 1991/92 and 1992/93, two years of good
rainfall with universal distribution of small free fertilizer
packs in 1998/99 and 1999/2000, poor rainfall and lower
fertilizer subsidies and production with widespread hunger
in 2000/01, 2001/02, and 2004/05, and good rainfall and a
large fertilizer subsidy in 2005/06.
Fertilizer use also rose impressively through the 1990s,
with an annual average of 6 percent growth in fertilizer use
on all crops and commercial and smallholder farms between
1984/85 and 2004/05 (SOAS et al. 2008). Starting in the
mid-1990s, private input suppliers took over an increasing
share of the fertilizer market from ADMARC and the Small-
holder Farmers Fertiliser Revolving Fund (SFFRFM), both
parastatals responsible for importing and distributing fertil-
izers to smallholder farmers. By the end of the 1990s private
input suppliers were responsible for more than 70 percent
of national fertilizer imports and for a large proportion of
sales to smallholders (SOAS et al. 2008). During the 2002/03
and 2003/04 crop seasons, 43 percent of smallholders sur-
veyed in a nationally representative survey purchased some
fertilizer (see table 17.1). Smallholders using fertilizer on
maize applied an average of 101 kilograms per hectare (see
table 17.1). Major parastatal involvement in fertilizer
imports for subsidized fertilizer sales has, however, affected
private sector sales and confidence in investment in imports
and retail systems by varying, and debated, amounts.
Widespread fertilizer use on maize produced by small-
holder farmers is constrained by two problems: prof-
itability and affordability. Unsubsidized fertilizer was not
generally profitable on maize produced for sale in Malawi
from the mid-1990s to the mid-2000s.4 It was, however,
more profitable on maize grown for households’ own con-
sumption, with a higher subjective valuation stemming
from farmers’ fears of the effects of a bad year on maize
purchase prices. For poorer farmers with this higher sub-
jective valuation, however, affordability of fertilizer
becomes a major problem. Liquidity presents substantial
difficulties for poor farm households, who on the one hand
face a “hungry gap” during the cropping period (when
farmers need to invest labor, seed, and other inputs in crop
production while food stocks from the previous season are
running low, and children are particularly susceptible to
sickness) and on the other very high borrowing costs and
an absence of low-cost financing services for inputs. Hun-
gry-gap problems at the livelihood level are exacerbated by
rural economy market effects (which depress wage rates
and asset prices and raise food prices). These problems are
widely recognized as very severe for poor rural households
292 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
in Malawi, causing major production and welfare prob-
lems in rural areas.5
Improving the profitability of fertilizer use in maize pro-
duction requires lower fertilizer prices (as a result of greater
efficiency in fertilizer supply and reduction in transport
costs for importation or distribution of a subsidy, or both),
higher maize prices, or greater efficiency in the use of fertil-
izer (raising the grain output to nitrogen ratio).6 Changes to
maize prices and improved efficiency of fertilizer use will
not, however, improve the affordability of fertilizer for large
numbers of poor rural households in Malawi. Making fer-
tilizer more affordable requires very substantial reductions
in fertilizer prices, the development of low-cost and accessi-
ble financial services, or both. Development of such finan-
cial services, however, requires that maize be profitable, that
smallholders have other sources of cash income that can be
used to repay fertilizer loans when the majority of the maize
they produce is for home consumption, and that very low-
cost systems be used for loan disbursement and recovery. All
these requirements are difficult to achieve in Malawi.
REVIEW OF INPUT SUBSIDIES
To understand the implementation, impacts, strengths, and
weaknesses of the Malawi Agricultural Input Subsidy
Program, it is helpful to examine broader historical and the-
oretical lessons on input subsidies’ implementation, perfor-
mance, and impacts.7
Wider experience with agricultural input subsidies
Large-scale (so-called universal) agricultural input subsidies
were a common and prominent feature of agricultural
development policies in poor rural economies from the
1960s to the 1980s. They were subsequently criticized,
however, as a major element in fiscally and economically
unsustainable policies that were inefficient, ineffective,
and expensive in Africa (see, for example, World Bank
1981). These policies distorted market incentives, blunted
competitiveness and farmer incentives, and undermined the
growth of private sector agricultural services. While subsi-
dized input systems may have seemed attractive to farmers
(in regard to the services that were supposed to be pro-
vided), theoretical difficulties with subsidy benefits (see
below) were compounded by diversion and inefficiency,
which often limited actual benefits to farmers.
Evaluations of the rate of return to alternative public
investments in Asia tend to rank input subsidies as fourth or
fifth, after investments in road infrastructure, agricultural
research and development, education, and often other types
of public investments (see, for example, Fan, Thorat, and
Rao 2004; Economist Intelligence Unit 2008). There are also
arguments, however, that while returns to agricultural input
subsidies were often low, they did yield substantial benefits
in some countries at certain times. Such arguments stress
the importance of differences between subsidies benefiting
fertilizer suppliers and those benefiting (poorer) farmers,
falling returns over time where subsidies are effective, and
the need for judicious (and changing) decisions on the scale
of different investments, recognizing trade-offs, comple-
mentarities, differences in the timing of returns, and poten-
tial diminishing (and sometimes increasing) marginal
returns across different investments (see Dorward et al.
2004; Djurfeldt et al. 2005; Timmer 1989, for Indonesia;
Fan, Gulati, and Thorat 2007, for India; and Dorward 2009).
Information on the performance of most of the recent
input subsidy programs in Africa is limited despite the very
substantial investments of public funds in these programs.
However, recent empirical evidence from Malawi and
Zambia shows that subsidies tend to be targeted dispropor-
tionately to better-off farmers compared with poor and
female-headed households, where affordability constraints
are most severe (Govereh et al. 2006; SOAS et al. 2008);
input subsidies have partially displaced commercial
fertilizer demand, which has hindered policy objectives to
promote sustainable development of commercial input dis-
tribution systems (Xu et al. 2009; Ricker-Gilbert, Jayne, and
Chirwa 2011); and the high costs of large-scale input subsi-
dies means that there are very substantial opportunity costs
in terms of forgone public investments, investments that, as
shown by the Asian experience discussed earlier, may have
greater long-term impacts on poverty reduction and agri-
cultural growth. Moreover, Kenya has achieved impressive
growth in fertilizer use on food crops based on strong com-
mercial demand for inputs after the liberalization of input
marketing and foreign exchange controls, without the use of
subsidies (chapter 17; Ariga et al. 2008; Ariga and Jayne
2009). The potential applicability of the Kenyan model, or
parts of it, to other Sub-Saharan African countries needs to
be considered, taking account of particular features of the
Kenyan situation and experience.
Dorward (2009), in a review of a number of input sub-
sidy programs across Africa, also notes apparent tendencies
for these programs to focus on production objectives and
producer welfare (largely ignoring potential benefits for
consumers and for wider pro-poor economic growth); for
poor integration of many programs with complementary
investments; and in some programs, for an unfortunate lack
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
of interest in improving effectiveness and efficiency. Two
further commonalities are a limited focus on replenishing
soil fertility and a strong prevalence of heavy subsidies
(50 percent to 100 percent subsidy rates) on rationed inputs
(Dorward 2009).
Theoretical benefits and costs of agricultural
input subsidies
Conventional arguments for subsidies in agricultural devel-
opment have focused on the promotion of increased
agricultural productivity through the adoption of new tech-
nologies (Ellis 1992). Reduced costs of subsidized inputs
increase the profitability of these technologies and reduce
the risks perceived by farmers in adopting them. Together
with credit and extension services, input subsidies were sup-
posed to help farmers implement, benefit from, and then,
with the withdrawal of the subsidy, fully fund efficient input
purchases and use themselves.
Supply and demand analysis of input subsidies shows
that because of deadweight losses, a subsidy can generate a
positive net economic return to a country only if it
addresses some market failure (Siamwalla and Valdes 1986).
This may occur when:
■ farmers’ private costs of working capital for input pur-
chase are greater than the social cost of capital,
■ farmers’ lack of knowledge about the benefits of inputs
means that their expectation of the production bene-
fits from input use are less than the benefits they will
gain,
■ there are learning costs with input use, meaning that ini-
tial farmer returns are low but will increase with experi-
ence (Ellis 1992; Crawford, Jayne, and Kelly 2006; Morris
et al. 2007), and
■ farmers’ risk assessment and aversion to investing work-
ing capital in input purchase and use are higher than
society’s risk assessment and aversion.
The size of the deadweight loss and the distribution of
benefits between consumers and producers also depend on
the elasticity of supply and demand. Demand or supply
inelasticity tends to be associated with smaller deadweight
losses. Inelastic demand is associated with larger shares of
consumer surplus benefits, while inelastic supply is associ-
ated with larger shares of producer benefits (Dorward
2009). Staple food markets in landlocked countries tend to
be associated with more inelastic demand by poor con-
sumers (where prices lie between export and import parity
293
 prices). Demand tends to be more elastic for cash crops,
particularly for cash crops that are exported.8
Subsidy inefficiencies also arise when part of the cost of
the subsidy goes to reducing the cost of production for pro-
duce that would be produced anyway; when subsidies bid
up demand and prices for land, labor, or inputs, and are
passed back to suppliers of these inputs,9 and when
rationing leads to opportunities for those controlling subsi-
dized inputs to demand payments for provision of subsi-
dized inputs. Another major concern with input subsides is
the extent of leakage and diversion away from their
intended use as a result of diversion between products,
diversion from intended beneficiaries to others within the
country, and cross-border leakage.
A final, crucial point is that the technical efficiency of
input use in generating additional agricultural production
is critical in determining deadweight losses, distribution of
benefits between producers and consumers, and wider eco-
nomic gains. This efficiency depends upon the quality and
appropriateness of the inputs to the product on which
they are used, the timing of the delivery of inputs to farm-
ers, the availability of complementary resources (for example,
seed and fertilizer together), agro-ecological conditions, and
farmers’ technical skill or competence in using the inputs.
This analysis suggests that large-scale input subsidies
should be focused on:
■ producers who are not using inputs because of market
failure,
■ crops and geographical areas for which increased input
use can induce a large supply shift (this may also require
complementary infrastructure and services for input
delivery, extension, and output markets), and
■ stimulating products with inelastic demand and supply,
particularly inelastic demand, among poor producers and
consumers (staple grain production tends to have these
characteristics in poor large or landlocked countries).
The analysis also suggests the importance of consumer
benefits in addition to (or rather than) producer benefits for
achieving economic and welfare gains from subsidies; sub-
sidy implementation that reduces deadweight losses and
rents from straight transfers, leakages, and high administra-
tive costs; and comparing distributional impacts and multi-
pliers from expenditure on input subsidies with alternative
(tax, subsidy, or other transfer) instruments for changing
income distribution and for stimulating growth.
The conclusions from this neoclassical supply and
demand analysis influenced conventional wisdom on a
294 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
number of difficulties with input subsidy programs: in con-
trolling costs, in achieving “exit strategies” after subsidy pro-
grams have become entrenched, in effective targeting of
input subsidies to particular farmer types, in ensuring inputs
are not overused, in controlling regressive benefits that favor
larger farmers who can access subsidized inputs, and in pre-
venting market distortions where parastatal involvement
crowds out private sector investment in input supply systems
and provides opportunities for corruption.
In recent years, however, some scholars and many gov-
ernment policy makers have departed from orthodox neo-
classical thinking on input (particularly fertilizer) subsidies.
Factors giving rise to this rethinking in Sub-Saharan Africa
include perceptions by some that liberalization policies have
failed to support sustainable intensification of staple food
crop production; political demands for fertilizer subsidies;
tensions among donors facing such demands; concerns
about declining soil fertility; and interest in using input sub-
sidies as an instrument for social protection policies and as
a means of promoting input market development.
On the basis of this analysis, Dorward (2009) suggests
that the following design and implementation features are
important if subsidy programs are to be effective and effi-
cient in stimulating increased productivity and broad-based
growth:
■ Large unit (or percent) subsidies on rationed supplies
targeted to credit constrained farmers to reduce input
affordability problems.
■ Targeting access to subsidized inputs for specific house-
hold types where input use is constrained by the market
failures that the program effectively addresses; where
these inputs can be used effectively and efficiently;10 and
where substantial political, economic, welfare, equity,
and administrative challenges in effective and low-cost
targeting can be overcome.
■ Rationing to control the costs of input subsidies with large
per unit or percent subsidies and limited secondary mar-
kets in which recipients sell subsidized inputs to others.
■ Encouragement of private sector input supply systems’ effi-
ciency and investments by economies of scale and by
competition in the sale of large volumes of inputs (espe-
cially in remote and previously poorer and less productive
areas and producers), with measures to limit uncertainty
and the diversion of suppliers’ focus to capturing subsi-
dized sales without developing retail systems.
■ Promotion of dynamic effects on pro-poor growth through
higher land and labor productivity in staple food pro-
duction, lower food prices, and higher producer incomes
 that facilitate wider nonagricultural development, mar-
ket thickening, and reduced coordination and transac-
tion costs and risks in poor rural economies.
■ Effective and efficient entitlement and distribution sys-
tems supporting targeting, rationing, supply system
development, control of secondary markets and leakages,
and cost control. A combination of paper vouchers (or
coupons),11 scratch cards, and electronic systems
(involving bank cards, electronic “smart” cards, mobile
phones, or some combination of the three) may be used
as evidence of entitlement. Different systems offer differ-
ent potential benefits and political, technical, adminis-
trative, and social challenges within communities and
households. Entitlements may be input specific or flexi-
ble with regard to inputs allowed and may have a fixed
value (with a variable top-up at redemption) or a vari-
able value (with a fixed top-up). There are also impor-
tant interactions between entitlement systems, secondary
markets, recipient choice (of inputs and suppliers), con-
trol of fraud and program costs, and gendered access to
and control of subsidized inputs within households.
■ Complementary investments, policies, and instruments
critical for subsidy effectiveness and efficiency, with bal-
anced investments in the subsidy program itself, research
and extension support, transport and communications
infrastructure, and efficient and stable output markets.12
■ Matching of political interests with more technical and
bureaucratic needs for cost control, limited leakages, tar-
geting, rationing, and private sector development.
Given the experience with subsidies in Africa discussed
earlier, implementing some of these features is a major
challenge.
Issues to consider in evaluating agricultural
input subsidies
The success of an input subsidy program has to be judged
against its objectives, and input subsidy programs can have
a wide range of possible objectives: wider (pro-poor) eco-
nomic growth, benefits for poor consumers from lower out-
put prices, national (or household) food self-sufficiency or
security, increased input adoption, increased efficiency of
input use, benefits for poor producers, input supply system
development, soil fertility replenishment, and political ben-
efits. Most, but not all, of these objectives can be mutually
complementary, depending on how a program is imple-
mented. The balance of program objectives, and their context,
should then determine the key design and implementation
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
elements of input subsidy programs as discussed earlier.
Figure 17.2 provides a conceptual framework that identifies
key variables and relationships affecting input subsidy pro-
gram impacts and guides this discussion of the Malawi
Agricultural Input Subsidy Program.
MALAWI’S 2005/06 TO 2008/09 EXPERIENCE
WITH AGRICULTURAL INPUT SUBSIDIES
Following severe food security difficulties in the early 2000s
and in line with election commitments, the government of
Malawi decided to implement a large national input subsidy
program for the 2005/06 growing season.13 The popular
program has been repeated and expanded in subsequent
seasons, building on core experience but also incorporating
modifications in components and implementation systems
from year to year. Core elements of the program common to
the different years have been its use of vouchers targeting
roughly 50 percent of farmers across the country for the
receipt of fertilizers for maize production, with further
vouchers for improved maize seeds and for fertilizers for
tobacco. The core objective of the program, which has been
refined over time, has been twofold: to increase resource-
poor smallholder farmers’ access to improved agricultural
inputs in order to achieve food self-sufficiency, and to raise
these farmers’ incomes through increased food and cash
crop production. The main features of the program across
the four years are summarized in table 17.2, and details of
its design, implementation and various achievements are
detailed in the following sections.
Program design and implementation
The 2005/06 program provided the foundation on which
subsequent input subsidy programs in Malawi have been
built. We therefore describe this program in more detail
before considering changes made in subsequent programs.
The objective of the program was to promote access to and
use of fertilizers in maize and tobacco production in order
to increase agricultural productivity and food security. Fer-
tilizer coupons were distributed to districts and within dis-
tricts to extension planning areas in two rounds. In the first
round, allocation was broadly in proportion to cropped
maize and tobacco areas. Coupons were distributed to dis-
tricts and Traditional Authorities (TAs) by the Ministry
of Agriculture and Food Security (MoAFS). TAs were
supposed to allocate coupons to Village Development Com-
mittees (VDCs), which were then supposed to identify
recipients to receive two coupons that they could redeem, at
295
 Figure 17.2 Conceptual Framework for Investigating Agricultural Input Subsidy Impacts

Other macro-
3. Effects on input supply economic
system 4. Effects on macro management
private sector importers and economy
large/ small distributors; fiscal balance Political
parastatals foreign exchange balance economy and
profits, cash flow, health, education, policy
confidence, volumes, prices, infrastructure, other processes
investment, innovations, agricultural spending
other services Global and
1. Input subsidy
implementation regional
Scale, cost, prices
modalities, timing,
targeting, rationing Staple food
market
policies
2a Effects on 2. Effects on 2b Effects on non-
recipients rural households recipients Previous
(different hholds and hhold season(s)
(different hholds and
members) events and
hhold members)
local and national, input access and use, farm and outcomes
input access and use,
farm and non-farm market and non- non-farm activities and
activities and productivity market relations: productivity, labour hire Weather
labour hire in/out (staple food, labour, in/out, crop purchases /
crop purchases/sales, cash, land, etc sales, income, food security,
Disease
income, food security, prices and flows) welfare, vulnerability
(HIV/AIDS,
welfare, vulnerability malaria, etc)

Other rural
economic
Other social protection and agricutural/rural development measures activities

Source: Adapted from SOAS et al. (2008).

Note: hholds = households.

Table 17.2 Summary of Malawian Agricultural Input Subsidy Programs, 2005/06 to 2008/09
Indicator 2005/06 2006/07 2007/08 2008/09
Fertilizer voucher distribution (mt equivalent) 166,156 200,128 216,000 195,369
Households receiving one or more fertilizer coupons (percent) — 54 59 65
Subsidized fertilizer (mt For maize) 108,986 152,989 192,976 182,309
For tobacco 22,402 21,699 23,578 19,969
Total subsidized Planned 137,006 150,000 170,000 170,000
fertilizer sales Actual 131,388 174,688 216,553 202,278
Redemption price (MK/50-kilogram bag) 950a 950 900 800
Approximate voucher value, (MK/bag) 1,750 2,480 3,299 7,951
Approximate subsidy (%) 64 72 79 91
Subsidized maize seed (mt) — 4,524 5,541 5,365
Hybrid seed (%) 0 61 53 84
Cotton seed (mt) 0 0 390 435
Legume seed (mt) 0 0 24 —
Cotton chemical vouchers 0 0 131,848 —
Total program Planned 5,100 7,500 11,500 19,480
cost (MK millions) Actual 7,200 12,729 16,346 39,847

Sources: Logistics Unit reports; Nakhumwa 2006; SOAS et al. 2008; MoAFS 2008; Dorward and Chirwa 2009a; MoAFS implemen-
tation guidelines; government of Malawi budget statistics; Dorward, Chirwa, and Slater 2010b; key informants.
Note: — = not available; mt = metric tons.
a. seed or fertilizer coupon (NSO 2008).

296 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
a reduced cash price, for any of four types of fertilizer. There
was considerable variation within areas in the criteria deter-
mining prioritization and selection of beneficiaries, in
numbers of people receiving coupons, and in numbers of
coupons received per recipient household. A second, sup-
plementary round of coupon allocation and distribution
was made later in the season. Under the program, 6,000
metric tons of open pollinated variety (OPV) maize seed
was also offered for sale without coupons, at a subsidized
price of MK 150 for a 3-kilogram bag, compared with the
market price of MK 500. Although the parastatals
ADMARC and SFFRFM were responsible for distributing
subsidized inputs, 48 percent of subsided fertilizer was sup-
plied by private sector importers.
Holders of coupons were entitled to redeem coupons
for fertilizer at the rate of one coupon and MK 950 for one
50-kilogram bag of 23:21:0 +4S or urea (“maize fertilizers”),
and at one coupon plus MK 1,450 per bag of Compound D
or CAN (“tobacco fertilizers”). These prices represented, on
average, a two-thirds subsidy to farmers on the market cost
of inputs. Coupons intended for different types of fertilizer
were not marked as such, and many coupons allocated for
“tobacco fertilizer” may have been used to buy “maize fertil-
izer” instead. Sales continued into January 2005, and in var-
ious areas were limited either by a shortage of fertilizer stock
or a shortage of coupons. In the latter case, supplementary
coupons were used in some areas, but unavailability of fer-
tilizer in time for it to be agronomically useful meant that
significant numbers of coupons were not used.
ADMARC/SFFRFM reported total sales of 131,803 metric
tons of subsidized fertilizer (representing 2.62 million
coupons).14
Malawi’s 2005/06 agricultural input subsidy program is
reported to have cost MK 7.2 billion, against a budget of MK
5.1 billion (SOAS et al. 2008). The reported program cost
excludes overhead costs for ADMARC and SFFRFM and
likely allows for only partial deduction of farmer payments
to ADMARC and SFFRFM for coupon redemption: these
payments amounted to a total of MK 2.7 billion.
Following the popularity of the 2005/06 program and
the perception of its success, the government decided to
implement the program in 2006/07 with a number of mod-
ifications (table 17.3). These included an increase in the
overall amount of maize fertilizers to be subsidized, a stan-
dard redemption price of MK 950 per bag for all fertilizer
types, improved coupon security (with differentiation by
fertilizer type), involvement of the Logistics Unit (a unit
largely funded by the U.K. Department for International
Development, which had played a major role in the logistics
of the nationwide starter pack and targeted input programs
from 1998/99 to 2004/05), involvement of several large
input supply companies in retail sales of subsidized fertilizer,
and use of maize seed vouchers that could be exchanged at a
wider range of outlets (including agro-dealers) for different
quantities of OPV or hybrid seeds.15 The seed component, a
portion of the Logistic Unit’s costs, and an independent

Table 17.3 Principal Changes in Subsidy Program Design and Implementation, 2005/06 to 2008/09
Year Subsidized inputs Voucher distribution system Voucher redemption systems Other system innovations
2005/06 Maize and tobacco fertilizers, District allocation by maize areas, Only through SFFRFM None
maize seed (OPV) distribution through TAs and ADMARC
2006/07 Maize and tobacco District allocation by maize areas, Fertilizers also through major Coupons specific to fertilizer
fertilizers, maize seed distribution varied, through local retailers; flexible maize seed type; fertilizer buy-back
(hybrid and OPV) government, TAs, VDCs, MoAFS vouchers through wide range system; involvement of
of seed retailers Logistics Unit
2007/08 Maize, tobacco, coffee, and District allocation by farm Fertilizers also through major Reduced copies of coupons;
tea fertilizers; maize seed households and areas, retailers; flexible maize and remote EPA premium;
(hybrid and OPV); legume distribution through legume seed vouchers through fertilizer buy-back system
seed (limited); cotton seed MoAFS and VDCs wide range of seed retailers;
and chemicals cotton inputs through ADDs
2008/09 Maize and tobacco fertilizers; District allocation by farm Fertilizers only through ADMARC Extra coupon security
maize seed (hybrid and households and areas; use of and SFFRFM; flexible maize and features and market
OPV); legume seed, cotton farm household register; open seed vouchers through wide monitoring; no remote EPA
seed, and chemicals; maize meetings for allocation and range of seed retailers; cotton premium; ADMARC
storage chemicals disbursement led by MoAFS inputs through ADDs computers for voucher
processing

Sources: Logistics Unit; Nakhumwa (2006); SOAS et al. (2008); MoAFS (2008); key informants; Dorward and Chirwa (2009a); MoAFS Implementation
guidelines; NSO (2008).
Note: EPA = Extension Planning Area.

CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 297
 program evaluation were funded by donors, who had not
directly financed any part of the 2005/06 program (other
than through budget support). Donors also funded a buy-
back scheme, which reduced the risks to government of
holding unsold stocks at the end of the year if private sales
led to lower-than-expected sales by ADMARC and
SFFRFM.
Planned and achieved subsidy sales and costs in 2006/07
(and other years) are shown in table 17.2. Supplementary
fertilizer voucher issues and the availability of fertilizer
for sales by private companies (which sold just under
30 percent of subsidized sales) together led to higher sales
volumes than budgeted. These, together with higher
prices than budgeted, led to significant budget overruns.
These problems were not faced in seed sales, where no
extra coupons were issued.
Growing experience with the program led to consolida-
tion in 2007/08 of many of the changes made in 2006/07,
together with further changes to extend the scope of the
program. Program objectives and beneficiary targeting cri-
teria were amended to give greater emphasis to concerns for
vulnerable households. Targeted quantities of subsidized
maize fertilizer and seed were again increased, to roughly
equal disbursements to the previous year. Changes were
made to coupon allocation systems between districts to pro-
vide greater weight to the number of farming households
(and less weight to crop areas), leading to an increasing pro-
portion of coupons allocated to the more densely populated
southern region, where levels of poverty and poverty inci-
dence are greatest.
Following problems in some areas in 2006/07, systems
for allocating and distributing coupons within districts were
also modified in 2007/08 to give less power to TAs and more
responsibility to MoAFS staff. In addition to the maize seed
vouchers provided with maize fertilizer coupons, extra
“flexible vouchers” allowed farmers to choose maize or
legume seeds (although legume seed supplies were very lim-
ited). A “remote area premium” was also introduced to pro-
vide incentives to private retailers to extend their networks
into areas with low coverage by private retailers; a premium
was provided on the subsidy paid to private sector retailers
for sales of subsidized fertilizers against identifiable vouch-
ers issued to beneficiaries in designated “remote areas” with
higher transport and distribution costs (with the vouchers
identifiable by their location code). Coupons for cotton seed
and chemicals were distributed through the MoAFS Divi-
sional Offices (ADDs).
Subsidized fertilizer volumes were again significantly
over budget in 2007/08; with higher-than-budgeted input
298 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
prices, program costs were 29 percent above the budget,
compared with 18 percent the previous year. However, pri-
vate sector subsidy sales were roughly the same as the previ-
ous year (increasing by only 6 percent, from 49,000 metric
tons to 52,000 metric tons), whereas parastatal sales
increased by approximately 30 percent, from 125,000 to
165,000 metric tons.
A number of further changes were made to the 2008/09
subsidy program. A farm household register compiled the
previous year was updated and used to list coupon alloca-
tions to individual beneficiaries in open village meetings led
by teams involving MoAFS and local government staff. An
attempt to print coupons in the government printer was fol-
lowed by a significant security breach, and vouchers for the
central and northern regions were then printed outside the
country with extra security features. The flexible maize and
legume seed voucher and cotton input systems were contin-
ued. Grain pesticides were also subsidized, and some subsi-
dized fertilizers were issued to tea and coffee farmers.
Although private retailers were initially involved in the sale
of subsidized fertilizers, this practice was discontinued at a
very late stage in the season. The program went massively
over budget, however, largely as a result of soaring interna-
tional fertilizer costs, but there was also an approximate
15 percent overrun in quantities of inputs subsidized.
Implementation achievements
Three aspects of Malawi’s achievements in implementing
its agricultural input subsidies are considered: program
scale, innovation and adaptation, and implementation
performance.
Program scale. Malawi’s agricultural input subsidy pro-
gram has grown each year since its creation and involves
complex and very significant logistical and organizational
challenges to tight deadlines. The major tasks of the pro-
gram are shown in figure 17.3. This summary is highly
simplified, however, and in practice a complex set of inter-
actions between various stakeholders is needed to complete
each task. In 2008/09 this process involved the selection of
more than 1.5 million fertilizer coupon beneficiaries from
more than 2.5 million farm households, printing and distri-
bution of 5.9 million coupons, and purchase and distribu-
tion of more than 3.4 million bags of fertilizer—all within
tight deadlines. Further challenges arise because farmers
served by the program are widely dispersed across the coun-
try (with a significant number being illiterate or semiliterate
and living in remote and poorly accessible areas), and
Figure 17.3 Major Tasks in Implementing Malawi’s Agricultural Input Subsidy Program
Planning and Farmer
budgeting registration
Input Secure coupon Coupon
purchase printing allocations
Coupon Beneficiary
distribution identification
Input distribution
(transport and
storage)
Coupon
issue
Coupon
redemption
Source: Authors.
because there is a constant threat of fraud or theft of program
commodities worth a total of approximately $220 million,
with each fertilizer coupon worth more than 10 percent of
annual household income for the more than 40 percent of the
population below the poverty line.
The implementation challenges should not be underesti-
mated, and it is significant that Malawi has been imple-
menting such activities at varying scales annually since the
inception of the starter pack program in 1998 (see below)
and has therefore built up both systems and considerable
expertise in these tasks.
Innovation and adaptation. The 2005/06 subsidy pro-
gram built on Malawi’s innovative experience in imple-
menting the starter pack and targeted input programs
(TIP). These programs involved large-scale registration and
targeting across the country; development of systems using
vouchers; coordination across different government, paras-
tatal, private sector, donor, and community stakeholders;
and substantial logistical challenges. The 2005/06 subsidy
program, however, involved a change in objectives (from an
emphasis on social protection and food security for vulner-
able households in the TIP to national food production and
self-sufficiency), an increase in the scale of subsidized
inputs (from approximately 50,000 metric tons of fertilizer
in 2004/05 to 130,000 in 2005/06), and the addition of
tobacco inputs and cash redemption of vouchers.
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Coordination
and control
Stakeholders
Farmers
MoAFS: HQ, LU, ADDs, DADOs, Ass, FAs
DCs, TAs, VDCs, Police, CSOs
Fertiliser importers, retailers
Seed producers/importers, retailers
ADMARC: HQ, districts, markets
SFFRFM: HQ, depots, markets
Transporters
Donors
Payments
and control
Following the experience of 2005/06, the government
has, with other stakeholders, implemented further innova-
tions to improve performance of the program and broaden
its impact. These changes emerged from formal and infor-
mal management and evaluation reviews and lesson learn-
ing within the government (formal internal evaluations
were conducted in 2006/07 and 2007/08); from discussions
with other stakeholders (donors, private sector fertilizer
importers, seed and fertilizer suppliers, a parliamentary
committee on agriculture, and civil society); from external
evaluations (commissioned by CISANet for 2005/06 and by
the government, DFID, and USAID for 2006/07 and by the
government and DFID for 2008/09); and shifts in policy
within a changing economic and political environment.
The major program modifications in 2006/07, 2007/08,
and 2008/09, summarized in table 17.3, concerned:
■ the extent and modalities of private sector involvement in
fertilizer imports, fertilizer sales, and seed sales, with a
buy-back scheme to reduce government stockholding
risks, a premium to stimulate private retail network
development in more remote areas in 2007/08, and exclu-
sion of the private sector from fertilizer sales in 2008/09;
■ recognition of the importance of including vulnerable
households among targeted beneficiaries, with an
increasing volume of inputs for maize production and
modified district/EPA allocation systems;
299
 ■ trialing of flexible vouchers for seed inputs and addition
of cotton inputs and grain storage pesticides;
■ introduction of beneficiary registration and more open
and more tightly managed beneficiary selection, voucher
distribution, and market monitoring systems;
■ coupon design, printing, security, and farmer redemp-
tion prices; and
■ sharing of the costs of some program components with
donors.
Implementation performance. Effectiveness and effi-
ciency of implementation can be assessed in terms of volume
of subsidized inputs disbursed, timing of subsidy sales and
supplier payments, targeted beneficiary access to inputs, and
cost. As shown in table 17.2, both planned and disbursed vol-
umes of subsidized inputs increased steadily between
2005/06 and 2008/09. Although fertilizer disbursement and
sales targets were not met in 2005/06, they were exceeded in
2006/07, 2007/08, and 2008/09 by 16 percent, 27 percent, and
19 percent, respectively. Exceeding these targets demonstrates
considerable success in meeting demand, but also suggests
difficulties in controlling disbursement and cost overruns.
Timing of subsidy sales is determined by the timing of
availability of inputs in markets and the issue of vouchers to
beneficiaries. Timing is critical for the effective use of seed
and fertilizer at the start of the agricultural season. For fertil-
izer, the timing of input availability depends upon timing of
tendering of input purchases and supplier deliveries to
depots, the staffing and stocking of input markets (for paras-
tatal sales), and subsidy redemption contracts with retailers
and their stocking and staffing of input sales points for private
sector sales. The timing of voucher issue depends on the tim-
ing of beneficiary registration, voucher allocations, voucher
printing, voucher distribution to districts, and district distri-
bution payments. Information on some of these variables is
given in table 17.4. In general, performance regarding earlier
award of seed and fertilizer contracts and earlier fertilizer
deliveries to depots and uplifts has improved over time. Infor-
mation on the timing of fertilizer sales is incomplete, but it
appears that despite some evidence of improvement between
2005/06 and 2006/07 (not shown in table 17.4), there has
been little improvement since then. It is particularly impor-
tant to increase sales by the end of November. These sales
were highest in 2008/09 but still only 30 percent of the total.
Receipt of seed vouchers by the Logistics Unit is determined
by the timing of sales and the speed of voucher processing by
seed suppliers; both were problematic in 2007/08.
Targeted beneficiary access to inputs is determined by
coupon allocation and issue and by the use of coupons,
300 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
which may be affected by the availability of subsidy inputs
in accessible markets and by any “tip” needed to redeem
coupons. Household surveys provide the only systematic
information available on these areas. Results from focus
group discussions and household surveys examining the
2006/07 and 2008/09 programs (SOAS et al. 2008; Dorward,
Chirwa, and Slater 2010a) suggest the following:
■ In 2008/09, 65 percent of farm households received one
or more fertilizer coupons, with an average of 1.5
coupons per household receiving coupons and of 1.1
coupons per household across all households (figures for
2006/07 were 54 percent, 1.7, and 1.0, respectively);
■ Targeting criteria were highly variable across different
administrative areas;16
■ Overall targeting recommendations were followed to
some extent in that there was a tendency for targeting
to reach households that are productive full-time
farmers. However, household survey data for the of
2006/07 and 2008/09 seasons indicate that coupons
were disproportionately targeted to households with
relatively large amounts of land or other assets, and (in
2006/07 but not in 2008/09) to male-headed house-
holds. Smaller proportions of fertilizer coupons were
given to households in the bottom half of the wealth
and income distribution;
■ In some areas, particularly the south and center, coupon
allocations were modified so that in 2008/09 just under
40 percent of households in these regions and 36 percent
nationally received one fertilizer coupon (rather than
fewer households receiving two coupons);
■ Open meetings for coupon allocation were introduced in
2008/09 and appear to have succeeded to some extent in
increasing the proportion of coupons and subsidized fer-
tilizer going to poorer households (Chirwa, Matita, and
Dorward 2010);
■ Key informants tended to underestimate the propor-
tion of households receiving subsidized inputs com-
pared with estimates provided in interviews with
households;
■ In 2006/07, 75 percent of ADMARC and private supplier
outlets and 100 percent of SFFRFM outlets were reported
to have suffered from frequent major queues; a similar
figure of 75 percent was reported across ADMARC and
SFFRFM outlets in 2008/09;
■ In both 2006/07 and 2008/09, household surveys indi-
cated that 5 percent of coupons were reported to be
accessed with some payment, with a median price of MK
1,000 in 2006/07 and of MK 2,000 in 2008/09.
 Table 17.4 Implementation Performance Indicators 2006/07–2008/09
Indicator 2006/07 2007/08 2008/09
Fertilizers
Tender awards for parastatal supplies Late August Mid-August End-July
Depot receipts end October (% of
parastatal total sales) 32 58 53
Depot receipts, end November
(% of parastatal total sales) 77 76 71
Outstanding payments, end Nov
(% and MK millions) 28 1,216 22 1,595 16 3,500
Outstanding payments, end Dec
(% and MK millions) 46 4,303 13 1,192 13 3,690
Outstanding payments, end Jan
(% and MK millions) 14 1,406 21 2,620 — 7,707
Uplifts dispatched as of end November
(% of parastatal total sales) 64 70 75
Total relocation transport costs
(MK millions) — 68.4 42.0
Finalization of retail fertilizer contracts Early Nov Mid/late Nov —
District voucher allocations Early Sept Oct 9 Sept 12
Voucher printing End Sept End Oct SR early Oct CR/NR early Nov
Voucher and list distribution to
districts completed Nov 7 Nov 3 Nov18
Sales end Nov (% of total season sales) 8 — 30
Sales end Dec (% of total season sales 74 — 68
SFFRFM/ADMARC voucher returns
end Dec (thousands) 0 101 175
SFFRFM/ADMARC voucher returns
end Jan (thousands) 111 720 1,057
Finalization of seed supply contracts Mid/late Nov Mid/late Nov Early Nov
Seed coupons in LU end Dec
(% of season sales) 27 4 6
Seed coupons in LU end Jan
(% of season sales) 74 18 22

Sources: Logistics Unit reports; Nakhumwa (2006); SOAS et al. (2008); MoAFS (2008); key informants.
Note: Data are not available for 2005/06. LU = Logistics Unit. — = not available.

■ In 2006/07 a “tip” was paid to retail market staff for
redemption of about 20 percent of fertilizer coupons,
with a mean price per bag of just over MK 1,000 (com-
pared with the official price of MK 950) and with no
significant overall differences between parastatal and
private sector suppliers. In 2008/09, 14 percent of fertil-
izer coupons were reported to require a “tip” for redemp-
tion, with a median “tip” of 200 MK, again giving a price
of MK 1,000 per bag.
There are considerable difficulties in determining the
extent to which fraud affects Malawi’s agricultural input
subsidy program. Fraud can arise in a number of ways—
through allocation of vouchers to nonexistent beneficiaries
(and their diversion to government staff, traditional leaders,
or politicians), direct allocation of vouchers to people who
do not satisfy beneficiary criteria, printing of extra or coun-
terfeit vouchers, and payment of “tips.”
Determining the extent of the fraud is rendered difficult
by the lack of formal and transparent audit systems cover-
ing the whole program and by discrepancies between
MoAFS and NSO estimates of the total number of farm
households in Malawi (MoAFS estimates of farm house-
holds were 33 percent above NSO estimates in 2006/07 and
47 percent above NSO estimates in 2008/09). SOAS et al.
(2008) concluded that there is insufficient evidence to sug-
gest widespread fraud, and that household survey estimates
of subsidized fertilizer access were broadly compatible with
the MoAFS farm household estimates, but there are numer-
ous anecdotal reports of fraud within the system. Although
Dorward, Chirwa, and Slater (2010b) suggest that NSO may
well underestimate the number of farm households in
Malawi, it seems unlikely that underestimate could be off by
a third, and there are risks (and anecdotal reports) of
increasing numbers of villages and of some “ghost villages,”
suggesting significant diversion of subsidized inputs away

CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 301
 from the intended beneficiaries. The discrepancy between
NSO and MoAFS estimates of the number of farm families
is being jointly examined by the NSO and MoAFS to resolve
this issue.
Regarding the extent to which counterfeit or nonstan-
dard vouchers (those with serial numbers outside the ranges
recorded by the Logistics Unit) have been accepted by dif-
ferent outlets, records for 2007/08 show that these (and sales
without vouchers) accounted for 27 percent of ADMARC/
SFFRFM sales and 3 percent of private retailer sales (LU
2009). Rapid return of vouchers to the Logistics Unit is
important for early identification of markets accepting
counterfeit or nonstandard vouchers. Private retailers gen-
erally return coupons quickly in order to receive payment,
but ADMARC and SFFRFM have been much slower at this
task although their voucher returns have improved over the
three years for which records are available (see table 17.4). A
major security breach in the printing of vouchers in 2008/09
led to reprinting of more secure vouchers for issue in two
regions.
Overall, the costs of Malawi’s subsidy program were over
budget and increasing from 2005/6 to 2008/9 (table 17.5),
due to a combination of increasing subsidy volumes and
large increases in fertilizer prices. Program costs were just
over 40 percent above the budget for 2005/06 and 2006/07,
nearly 50 percent over budget in 2007/08, and 90 percent
over budget in 2008/09. Program costs rose from just over
60 percent of the MoAFS budget in 2006/07 and 2007/08,
rising to 74 percent in 2008/09, when the program
accounted for over 15 percent of the total national budget.
It is, however, important to note that for the 2009/10 pro-
gram, actual costs were 21 percent below budget and costs
fell back sharply, by 41 percent, because of both a halving of
fertilizer prices and a 20 percent reduction in the amount of
fertilizer disbursed, which was almost exactly on budget. As
a result program costs fell back to 7 percent of the total
national budget.
Data on estimated per unit fertilizer costs and on total
program costs, excluding ADMARC overhead costs, are also
given in table 17.5. As shown in the table, fertilizer prices
and transport costs rose from 2005/06 to 2008/09. The esti-
mated per unit fertilizer cost increases from 2005/06 to
2006/07 in Malawi (25 percent) are higher than would be
expected given that international prices were static over
the same period, but from 2006/07 to 2007/08 the price
increase (22 percent) was markedly lower than the increase
in international prices, which rose by around 50 percent or
more, so the overall cost increase in Malawi over 2005/06 to
2007/08 was in line with international price increases.
Fertilizer cost increases from 2007/08 to 2008/09 also appear
to have been roughly in line with increases in international
price increases over the same period (about 125 percent).

Table 17.5 Fertilizer and Program Costs in Malawi 2005/06–2009/10

2005/06 2006/07 2007/08 2008/09 2009/10
Planned Actual Planned Actual Planned Actual Planned Actual Actual
Fertilizer cost ($/mt):
Parastatal: delivered
at depots — — — 454 — 555a — 1,204b 575
Parastatal: transport, etc. — — — 36 — 45 — 46 39
Parastatal: total — 393 — 490 — 600 — 1,250 614
Private retailers: total — ... — 490 — 612 — ... ...
Average all suppliers — 393 — 490 — 590 — 1,250 614
Program costs:
Malawi government 36.4 51.4 51.4 81.4 73.6 109.6 127 227.7 137.6
Donors ($ millions) 0 0 12.5 9.5 5.7 7.1 12.1 37.8 17.5
Total ($ millions) 36.4 51.4 63.9 90.9 79.3 116.8 139.1 265.4 155.1
Net of farmer payments — 32 — 73.9 — 95.4 — 242.3 143.7
Total as % MoAFS budget — — 43 61 51 61 61 74 —
Total as % national budget 4.3 5.6 5.4 8.4 6.7 8.9 8.5 16.2 6.7
Total as % of GDP — 2.1 3.1 — 3.4 — 6.6 —

Sources: Logistics Unit; Nakhumwa (2006); SOAS et al. (2008); MoAFS (2008); Dorward and Chirwa (2009a); Government of Malawi budget
statistics; key informants.
Note: The 2005/06 fertilizer costs may also include some seed and coupon production/distribution costs. Parastatal transport etc., costs exclude
ADMARC overheads. Program costs exclude buy-back carried forward. — = not available. ... = not applicable.
a. Excluding costs of buy-back brought forward.
b. Including costs of buy-back brought forward.

302 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Marked monthly variation in international fuel prices from
mid- to late 2006 and 2007 makes it difficult to calculate
equivalent figures for transport costs.
Aside from the problems of high fertilizer prices in
2008/09, the 2006/07 to 2008/09 programs faced major
challenges in controlling the volume of subsidized fertil-
izer disbursed. Three alternative (and complementary)
approaches to limiting the volume may be used: control-
ling the number of coupons issued, controlling the physi-
cal stock of fertilizer available, and controlling sales of
fertilizer by closing further sales once the total budget
quantity has been sold. In principle, the first option is the
best approach, although it is undermined by counterfeit
coupons and by any high-level political pressure that may
demand extra coupon issues. Control of physical stock of
fertilizer (the method used in 2009/10) is difficult if the
private sector is involved in retail subsidy, and may result
in genuine beneficiaries being denied the opportunity to
redeem genuine coupons. Closure of the program once
target sales have been achieved suffers from the latter dis-
advantage, and in addition requires timely reporting and
monitoring of sales.
The rising costs of the subsidy program over the period
were met by increasing budgetary allocations to the MoAFS,
and did not crowd out other MoAFS activities in terms of
actually cutting budgetary allocations to them. However, the
opportunity cost of the program is an issue in terms of for-
gone investments that could have been achieved with those
funds. The program also consumes very large amounts of
staff time and other resources, for people must be diverted
from other activities to manage and implement the subsidy
program in the critical time before and at the start of the
cropping season. Similarly, while financial resources allo-
cated to the subsidy program have grown dramatically since
the start of the program, financial resources allocated to
other activities have remained largely static or shown only
small increases, posing severe challenges to other essential
research and extension activities of the Ministry.
Input supply impacts
Malawi’s subsidy program has had major, and mixed,
impacts on private sector input suppliers. These effects have
to be considered separately for fertilizer importers, fertilizer
retailers, seed suppliers, and seed retailers. For retailers of
seed and fertilizer, it is important to distinguish between
small independent agro-dealers on one hand and retail out-
lets of larger companies involved in importation and both
wholesale and retail sales on the other.
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Fertilizer importers have been responsible for generally
increasing proportions and volumes of government subsidy
sales, with particularly large volumes in 2008/09 (table 17.6).
Importers have clearly benefited from their growing share in
imports, although they have faced some difficulties from
exposure to foreign exchange losses caused by delays in pay-
ments in local currency. Insofar as currency risks are fac-
tored into tender margins, these raise fertilizer costs for
government. Some importers have expressed concerns
about increasing competition in import tenders from new
players without proper qualification criteria, leading to the
award of some tenders to suppliers who are unable to
deliver. If the award of such tenders leads to late cancellation
and reordering at short notice, they may also raise fertilizer
costs for the government.
Maize seed suppliers have also benefited from significant
growth in sales over the life of the program (see table 17.2).
Government and the seed suppliers association negotiate
prices, which involves a difficult balance between competi-
tion and coordination in supply. The subsidy program
affects retail sales through three processes, each of which has
a different effect on whether retail outlets sell subsidized
inputs. These are set out in table 17.7.
Displacement of commercial sales occurs when a farmer
chooses not to buy an input received on subsidy when he
or she would have bought it commercially if the subsidy
had not been available. This affects private retail outlets
irrespective of their participation as subsidy retailers. Dis-
placement is difficult to estimate because even without sub-
sidies, farmers’ commercial purchases change from year to
year with changes in input prices, output prices, and their
access to seasonal finance. Input suppliers appear to be very
concerned about losses of fertilizer sales through displace-
ment if these are not counteracted by gains in subsidized
sales and customers from participation in the subsidy
scheme. Displacement rates have been estimated from
examination of changes in aggregate sales (SOAS et al. 2008)
Table 17.6 Private Sector Involvement in
Subsidized Fertilizer Sales
Private sector
involvement 2005/06 2006/07 2007/08 2008/09
Subsidized tender
deliveries (mt) 70,000 99,386 97,845 162,840
Subsidized tender
deliveries (%) 48 72 71 88
Retail sales (%) 0 28 24 0
Source: Logistics Unit; SOAS et al. 2008; Dorward and Chirwa 2009a.
303
 Table 17.7 Impacts of Subsidy Program on Seed and Fertilizer Private Retail Outlets
Processes by which the subsidy
program affects retail sales Participating retail outlets
Displacement of commercial sales Loss of sales
by subsidy sales
Sales of subsidized inputs Gain in sales
Gain/loss of customers going to outlets Gain in sales
to redeem their subsidy vouchers
General increases in demand as a result Gain in sales
of program-induced growth and income/
cash gains in previous season
Private fertilizer retailers Retail chains in 2006/07 and 2007/08
Private seeds retailers Agro-dealers and retail chains from
2006/07 to 2008/09
Source: Authors.
and from panel data analysis of farmer purchases (Ricker-
Gilbert, Jayne, and Chirwa 2011).
Displacement estimates from examination of changes in
aggregate sales were 20–30 percent in 2005/06 and 30–40
percent in 2006/07, with displacement for tobacco fertilizers
higher than that for maize fertilizers (SOAS et al. 2008). Dis-
placement estimated from panel data analysis of farmer
purchases was 23 percent for 2006/07 (Ricker-Gilbert, Jayne,
and Chirwa 2011) and 3 percent for 2008/09 (Ricker-Gilbert
and Jayne 2010), but in all years some further displacement
may be expected if some subsidized fertilizers are not
received by smallholders. Estimating displacement from
aggregate fertilizer sales for 2007/08 and 2008/09 has not
been possible due to lack of data on aggregate commercial
sales. Table 17.8 shows incremental fertilizer use estimates
for 2005/06 and 2006/07 and predictions for 2007/08 and
2008/09, assuming similar implementation in these years.17
Displacement of maize seed sales appears to be much lower,
with strong growth in commercial seed sales in 2006/07.
As table 17.7 shows, however, overall impacts of the sub-
sidy program on input sales depend not only on displace-
ment effects but also on the impact of participation or
exclusion in the program on subsidy sales and on customers
visiting the outlet. The last two lines of the table identify the
status of agro-dealers and retail outlets for larger companies
with regard to subsidized fertilizer and seed sales. Notably,
both reported a significant increase in sales in 2006/07 when
they were able to participate in subsidized seed sales and in
subsidized fertilizer sales (only retail outlets for larger com-
panies) (Kelly, Boughton, and Lenski 2010). Conversely, the
exclusion of the private sector from all retail subsidy sales in
2005/06 led to a substantial drop in reported fertilizer sales
from all retail outlets. Sales recovered in 2006/07 and
304 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Excluded retail outlets
Loss of sales
No effect
Loss of sales
Gain in sales
Agro-dealers in all years, retail chains
in 2005/06 and 2008/09
Agro-dealers and retail chains in 2005/06
only
2007/08 for the larger importers with retail outlets with
their inclusion in retail subsidy sales, but they again
reported declines in retail outlet fertilizer sales when they
were excluded from the program in 2008/09 (Kelly,
Boughton, and Lenski 2010). Small agro-dealers had been
excluded from retail sales of fertilizer subsidies during all
four seasons of the program.
Maize market impacts
The input subsidy program may affect maize markets in a
number of ways. We identify four potential impacts:
■ Direct impact through increased supply of maize for sale
and reduced demand for purchases by net surplus and
deficit farmers;
■ Indirect impacts as a result of policy changes influenced
by the subsidy;
■ In the longer term, if the subsidy program leads to rising
incomes, demand for maize should increase as a result of
consumption by both humans and livestock;
■ Finally, if the net effect of these impacts is to lower (or
raise) maize prices, then a supply response to increase (or
reduce) resources allocated to maize production should
be expected.
These impacts arise in the context of wider changes in
production (as a result of seasonal weather), in policies, in
regional and national maize markets, and in urban and
rural incomes (as a result of other processes of livelihood
change and growth). Although data and analytical limita-
tions make it difficult to tease out these different influences,
Malawi does have good information on maize prices. As
 Table 17.8 Estimated Incremental Fertilizer Sales and Maize Production, 2005/06–2008/09
Indicator 2005/06 2006/07 2007/08 2008/09
Incremental fertilizer sales as % of subsidy sales 70–80 60–70 60–70 90
Incremental fertilizer use (mt) 98,541 113,547 140,760 181,800
Incremental seed use (mt) OPV 3,000 1,764 2,604 833
Hybrid 0 2,760 2,937 4,532
Yield response as % of 2008/09 estimate 80 100 70 100
Subsidy program incremental maize Medium estimate 406,348 647,474 566,235 968,900
production estimates (mt) Above 2002/03 and 2003/04 273,609 514,735 433,496 836,161
High estimate: +20 percent 487,618 776,969 679,482 1,162,800
Above 2002/03 and 2003/04 328,332 617,683 520,196 1,003,514
Low estimate: –20 percent 325,078 517,979 452,988 775,200
Above 2002/03 and 2003/04 218,887 411,788 346,797 669,009
National crop production estimates Increment above 2002/03
and 2003/04 (mt) 975,262 1,698,956 1,031,938 2,031,816
Net maize exports in following year (exports minus imports, mt) –78,491 224,972 –101,027 –50,398
Source: Incremental fertilizer and seed sales figures are from Dorward and Chirwa (2009a). Figures for yield response and incremental seed impact
for 2008/09 are from Dorward and Chirwa (2010). National crop estimates are from MoAFS. Net maize exports are from Jayne et al. (2010).
Note: Production seasons 2002/03 and 2003/04 are considered nondrought years for presubsidy comparisons, although targeted input subsidies of
35,000 and 22,000 metric tons of fertilizer were provided in these years, and a 10 percent displacement is assumed for these years. OPV sales for
2005/06 are estimated to be 50 percent of budgeted sales.

noted earlier, prices have varied widely in the past, and, as
shown in figure 17.4, variability has continued in the period
of subsidy implementation—indeed, maize prices reached
historic highs in early 2009.
High maize prices would not be expected given the large
maize production estimates in each subsidy year—rather,
low prices would be expected.18 Low prices were observed in
the 2006/07 marketing season (following the 2005/06 sub-
sidy) and initially in 2007/08 (following the 2006/07 sub-
sidy). In the latter year, however, prices rose toward the end
of the season, so that the maximum monthly price within the
year was high, although substantially larger estimated pro-
duction as compared with the previous year should have led
to lower prices. Prices following the 2007/08 subsidy were
even higher, with average annual prices exceeding those of
the 2001/02 and 2005/06 famine years,19 but these high
prices did not lead to any reports of widespread suffering
and distress such as those experienced in previous years with
equivalent prices. Prices in the first half of the 2009/10 sea-
son (following the 2008/09 subsidy) were considerably lower
than in the previous year, although the estimated production
was still very high by historical standards (figure 17.5).
A number of explanations for this pattern of prices and
estimated production may be put forward:
■ In 2007/08, roughly 330,000 metric tons of maize were
exported. In 2009/10, 130,000 tons were purchased by
government for the strategic grain reserve (SGR), and a
further 100,000 tons are estimated to have been bought
and held in storage by private traders (FEWSNet 2009).
Exports and purchases of stocks to carry over to the fol-
lowing season would reduce maize volumes (this would
not be the case for 2009/10 SGR and other stock pur-
chases subsequently sold later in the 2009/10 season).
The effect of this on figure 17.5 would be to shift the
2006/07 and 2008/09 subsidy figures to the left by around
300,000 and 100,000 metric tons, respectively.20 This
does not, however, bring the three later subsidy years
anywhere near the pattern of the 1993/94, 1995/96 to
2004/05, and 2006/07 seasons. Indeed, there were reports
of maize imports of roughly 50,000 tons from Mozam-
bique and 36,000 tons from South Africa during the
2008/09 season (FEWSNet 2009; South African Revenue
Service 2009).
■ Rising real incomes alongside falling poverty rates and
rising population may lead to rising national demand,
which would cause the 1995/96 to 2004/05 demand to
increase over time—but not as suddenly and dramati-
cally as shown in figure 17.5. Such a trend is compatible
with the lack of distress in later years despite high prices.
■ Storage losses may be rising as a result of increasing pro-
duction of hybrid maize promoted by the 2006/07 and
subsequent subsidy programs. However, 2009 household
survey results suggest that storage losses are not particu-
larly high, with 50 percent of respondents reporting no
losses in the 2007/08 and 2008/09 storage years and only
a little over 20 percent reporting high losses (Dorward
and Chirwa 2009b). Mangisoni (2010) also reports rela-
tively low storage losses, in the range of 12 percent over a
10-month period.

CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 305
 Figure 17.4 Mean Annual Maize Price in Malawi, by Marketing Season

0.40 Subsidy years

0.35

Maize price at 1990 US$/kg

0.30

0.25

0.20

0.15

0.10

0.05

0
2

0
/9

/9

/9

/9

00

/0

/0

/0

/0

/1
91

93

95

97

01

03

05

07

09
/2
19

19

19

19

99

20

20

20

20

20
19
Source: MoAFS.
Note: Maize prices are simple averages across all markets obtained from weekly surveys, all deflated to 1990 prices.

Figure 17.5 Peak Monthly Maize Prices in Malawi by Estimated Maize Supply per Capita, by Season

0.50

0.45 2000/01
Maximum maize price per kilogram

1994/95 2007/08
2004/05
0.40
1996/97
1997/98
(in 2005 U.S.$)

0.35 1993/94

0.30 2006/07
2001/02 2008/09
0.25 1995/96
1998/99
0.20 2003/04
2002/03 1999/2000
0.15 2005/06

0.10
0.10 0.15 0.20 0.25 0.30
Per capita maize crop estimates plus net imports (metric tons)

1993/94–2004/05 2005/06–2008/09

Source: Calculated from MoAFS annual crop production estimates and weekly maize price data with export-import data from Jayne et al. 2010, Minot 2009,
and FAO 2009.
Note: Estimated maize supply = crop estimate plus exports – imports. Labels show production season.

■ Higher welfare and real incomes following the 2005/06 ■ Finally, national maize production following the imple-
harvest and low maize prices led to greater retention and mentation of the subsidy program could have been over-
consumption of the 2006/07 harvest and hence a thinner estimated. Although there are no clear changes in
and tighter market. methodology in the last few years, the method appears to
■ Changes in informal cross-border flows could also have rely substantially on field workers’ subjective estimates of
occurred (Jayne et al. 2010). crop area and yield, which may be affected by the very

306 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
 substantial involvement of field workers in the subsidy
program. Production impacts of the subsidy program are
discussed in more detail below.
None of these explanations (except possibly the last) can
fully explain the high prices despite the high estimated pro-
duction following the 2006/07, 2007/08, and 2008/09 sub-
sidy programs. To illustrate this, figure 17.5 plots maximum
monthly price against estimated per capita net maize sup-
ply. From the 1993/94 to 2004/05 production seasons, there
was a roughly downward sloping relationship, with high
prices following years of low supply and low prices in years
following high supply (although the 2002/03 and 2003/04
seasons do not fit this relationship). Prices following the
first (2005/06) subsidy fit this pattern. The three subsequent
seasons, however, when prices were high despite high esti-
mates of production, do not fit this pattern.
Two clear and important conclusions emerge from this
analysis. First, in three out of four years, the subsidy pro-
gram did not lead to lower market prices for maize. Second,
the subsidy program has not led to increases in maize sup-
ply as large as those suggested by increases in maize crop
estimates from 2006/07 onward, particularly in 2007/08 and
2008/09, when prices appeared to be very high compared
with estimated supplies. We therefore consider estimates of
the subsidy programs impacts on production.
Production impacts
The major stated objectives of Malawi’s subsidy program
have been to achieve food self-sufficiency and to increase
the incomes of resource-poor households through
increased food and cash crop production. Increased pro-
duction is therefore critical to achievement of program
objectives. This results from incremental use of inputs
(mainly fertilizers and seeds) leading to increased yields,
with yield responses to these inputs dependent upon the
weather and the efficiency of input use and of crop produc-
tion. Estimated incremental fertilizer sales were discussed
earlier in terms of the effects of displacement on input sup-
ply markets. Incremental fertilizer sales are also important
for estimating the incremental production effects of the
program, with responses to fertilizer depending upon rain-
fall, crop variety, and management (including timing of
planting, weeding, and timing and methods of fertilizer
application), and soil fertility.
SOAS et al. (2008) and Dorward and Chirwa (2009a) cal-
culate estimated incremental production for 2005/06 to
2007/08 using a range of 12 to 18 kilograms of grain per
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
kilogram of nitrogen. Results from crop-cutting survey esti-
mates in the 2008/09 crop year demonstrate that substantial
problems in obtaining precise estimates of crop responses21
make it difficult to obtain precise estimates of incremental
production from the subsidy program. They do, however,
support the broad response of 12 to 18 kilograms of grain
per kilogram of fertilizer, with 15 kilograms per hectare a
reasonable “medium expectation.” Using different estimates
of incremental fertilizer use and of yield responses gives dif-
ferent estimates of incremental production as a result of the
input subsidy program. Table 17.8 sets out such estimates by
year of implementation.
For each year estimated, incremental fertilizer use is mul-
tiplied by a grain-to-nitrogen response ratio adjusted to
reflect differing conditions in subsidy implementation
between years (good weather but very little hybrid seed
in 2005/06 compared with 2008/09; similar conditions in
2006/07 but a bit less hybrid seed; late fertilizer delivery
in the southern region and only slightly more hybrid seed in
2007/08). This is added to a yield gain from subsidized
hybrid seed separate from fertilizer to arrive at an estimate
of total incremental maize production from the subsidy.22
High and low estimates are, respectively, 20 percent above
and below the 2008/09 medium estimate (which averaged
15 kilograms of grain per kilogram of fertilizer across
hybrid and local seed plots).
While the estimates in table 17.8 are necessarily approxi-
mate, indeed indicative, as noted above, they nevertheless
demonstrate several important points:
■ Incremental production is very sensitive to yield
responses to inputs (hybrid seed and fertilizers) and the
potential is therefore considerable for raising yields and
yield responses with good subsidy program and crop
management—with early subsidy sales, planting and fer-
tilizer application, high plant populations, and greater
use of organic matter, for example.
■ Not explicitly shown in table 17.8 is the importance of
hybrid seed in raising yield responses to fertilizer.
Increasing hybrid seed sales (subsidized or unsubsidized)
is therefore another potential way of increasing subsidy
impacts on incremental production.23
■ Incremental production estimates are considerable, and
they grow over the life of the program as a result of the
increasing volume of incremental fertilizer use and
increasing supply of hybrid seed.
■ Incremental production estimates are, however, consid-
erably less than the production increases estimated in the
national crop estimates for maize production since the
307
 start of the subsidy program, and these differences are
Table 17.9 Trends in Macroeconomic Performance
too large to be explained by upward revision of the yield Indicators, 2005–09
response to fertilizer.24 (percent)
■ Differences in production between presubsidy and sub-
Indicator 2005 2006 2007 2008 2009
sidy years as estimated above are more compatible with
Real GDP growth 3.3 6.7 8.6 9.7 6.9
price differences between these years as shown in figure Inflation 15.4 13.9 8.0 8.7 10.1
17.5 (for example the export of 330,000 metric tons fol- Deficit/GDP
lowing the 2006/07 harvest would more than cancel out ratio (budget) 2.6 1.5 1.8 1.9 3.7
Deficit/GDP
the increased subsidy impact that year, compared with ratio (actual) 0.4 1.4 4.0 6.3 8.0
2005/06). The very high prices in 2007/08 remain a puzzle
Source: Reserve Bank of Malawi 2010.
but may be explained by the subsidy program’s incre-
mental production being insufficient to counteract pro-
duction losses from adverse conditions affecting all maize production, and improved macroeconomic management
in some parts of the country (late subsidized input delivery and conditions were other important contributors).
and local events such as flooding and drought spells). Improved macroeconomic management, together with
budgetary support from donors, was also undoubtedly
important in enabling the government to finance such a
Macroeconomic impacts
large program.
The large size of Malawi’s subsidy program could be The situation was, changing, however, as increasing vol-
expected to have macroeconomic impacts. As a proportion umes and increasing prices in subsequent years led to very
of total government expenditure, the subsidy increased high cost overruns. At the same time, Malawi’s economy was
from 5.6 percent in 2005/06 to 8.4 percent in 2006/07 to 8.9 facing a number of internal and external pressures that led to
percent in 2007/08. With very large increases in fertilizer adverse changes in macroeconomic indicators. The subsidy
prices and costs for 2008/09, actual expenditure on the sub- program both contributed to and was affected by these
sidy rose to 16.2 percent of total government expenditure macroeconomic changes, this time adversely (other macro-
(see table 17.5). As a proportion of gross domestic product economic pressures were very high government expenditure
(GDP), subsidy program costs rose from 2.1 percent in and import costs for the subsidy program with high fuel
2005/06 to 3.4 percent in 2007/08 and to 6.6 percent in costs, high maize prices, other government expenditures, and
2008/09 (excluding remittance by ADMARC and SFFRFM lower tobacco prices), although the incremental maize pro-
of the farmer’s redemption price to government). (As noted duction from the subsidy program should have exerted a
earlier and shown in table 17.5, program costs have subse- downward influence on maize prices. In a fixed exchange
quently fallen back markedly.) rate environment, these pressures contributed to a foreign
On the positive side, estimates of GDP growth have been exchange crisis in Malawi in November and December 2009.
significantly affected by large increases in estimated maize Very high budgetary and foreign exchange allocations to
production since the implementation of the subsidy pro- the subsidy program also reduced funding available to activ-
gram. Estimates of incremental production attributable to ities such as health, education, and infrastructure develop-
the subsidy program are not as high but are nevertheless ment. It is clear that the 2008/9 level of spending on the pro-
very large, and incremental maize production and increases gram was not sustainable, and as noted earlier the
in land and labor productivity in maize production attrib- government is addressing this: although the very high fertil-
utable to the program should have had a significant positive izer prices in the 2008/09 season were a temporary phe-
impact on GDP growth. nomenon, the government has committed itself to control-
No evidence of negative macroeconomic impacts was ling costs by limiting the volume of subsidized fertilizers in
found in the subsidy’s first two years (SOAS et al. 2008). future years. It has also restricted the subsidy to inputs for
Important contributors to GDP were sound macroeconomic the production of only maize.
management; improving macroeconomic indicators includ-
ing growth, inflation, and government deficit (table 17.9);
Economic returns
and increased growth across the economy at that time; and
the subsidy program itself was a contributor to that Economic returns to Malawi’s subsidy program depend
growth (good tobacco prices, good weather for agricultural upon the economic price of maize, the price of inputs, and

308 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
production responses to increased input use. Producer ben-
efit-cost ratios estimated for the 2006/07 program showed
that the net economic return to the project is very sensitive
to maize prices and the production response, and, with rea-
sonable variation in assumptions, these ratios range from
0.81 to 1.30, with a mid estimate of 1.06 (table 17.10).
Adjustments to this analysis using estimated maize and fer-
tilizer prices for other program years suggest that both the
2005/06 and 2007/08 programs should have yielded equiva-
lent or higher returns. However, the very high fertilizer
prices that prevailed when fertilizers were being purchased
for the 2008/09 program adversely affected returns in
2008/09, despite good weather and yields and high maize
prices (although these did offset the effects of high fertilizer
prices to some extent).
Fiscal efficiency estimates (net economic benefit per
unit of fiscal investment) show a similar pattern to eco-
nomic returns, but in addition these are (negatively)
affected by high rates of displacement of unsubsidized sales
by subsidized sales (displacement lowers the net benefit of
subsidized sales). Key conclusions from the benefit-to-cost
and fiscal efficiency analyses are that economic returns are
highly sensitive to the yield response to fertilizer (as dis-
cussed earlier under production impacts); fiscal returns are
highly sensitive to displacement rates; and with good pro-
gram implementation and good (but achievable) yield
responses to fertilizer, the program can be a very good
investment. It is therefore critical that the program design
and implementation deliver low displacement and high
responses to inputs.
Growth and poverty reduction impacts
While the producer benefit-cost and fiscal efficiency analy-
ses can yield valuable information about the efficiency of
the subsidy program, they can be misleading when examin-
ing the contributions of the program to poverty reduction,
economic growth, and food security. Understanding the full
economic benefits of the program requires consideration of
the direct effects of the program on subsidy recipients and
of the different ways that these effects subsequently work
through their own and others’ livelihoods and the rural
economy. Because the program is large, it is very important
for the wider market effects of the intervention to be prop-
erly recognized.
Figure 17.6 shows three possible uses of the subsidy by
subsidy recipients: reselling of coupons or of subsidized
inputs, incremental use of the inputs in production, or use
of the inputs with displacement of purchase of unsubsi-
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
dized inputs. These should lead to two main types of direct
benefit for recipients: immediate income transfers (from
reselling of coupons or subsidized inputs or from reduced
expenditure on inputs as a result of displacement of unsub-
sidized purchases by cheaper, subsidized purchases), or
incremental production at harvest if the inputs are used on
farm. If poorer households sell their coupon(s), then imme-
diate income and welfare gains should also be accompanied
by an easing of short-term seasonal cash constraints. This
easing may reduce the extent to which they have to hire out
their own labor (as casual ganyu labor) to obtain food, thus
allowing them to work more on their own farms, both
increasing end-of-season yields and reducing the supply of
labor into the local labor market. At the same time, if less
poor households obtain cheaper inputs directly from the
program or buy those inputs from subsidy recipients, then
this increase in income should increase their demand for
hired-on farm labor or for local goods and services. The
result should be a tightening of both demand and supply in
the local labor market, and a consequent rise in wages, to
the benefit of poorer households.
Further direct and indirect benefits may also accrue in
subsequent seasons. At the end of the season, higher maize
production should result from incremental fertilizer use
and increased crop labor inputs, increasing households’
incomes. Higher production should also depress maize
prices. Higher household maize stocks and lower maize
prices carried forward into the following season should
again benefit poorer households, reducing their seasonal
cash flow constraints and their need to hire out their labor,
so that they can again work more on their own farms. This
will again tighten the labor market. Higher incomes from
higher wages should stimulate demand for nonfarm goods
and services, with spin-off benefits and multipliers in the
local economy.
There is considerable empirical support from Asia and
from Africa for the importance of some of these processes
and of indirect effects of agricultural growth on wider eco-
nomic growth. Hazell and Rosegrant (2000) show that the
indirect effects of increased agricultural production in the
green revolution in Asia were the major process driving pro-
poor growth in the second half of the 20th century. There is
also a large body of literature on agricultural growth “mul-
tipliers” in Africa, with estimates that vary from around 1.5
to over 2.0 (a multiplier of 1.5 indicates that $1.00 of extra
income from agricultural production results in further
income growth of $0.50) (Hazell and Hojjati 1995; Reardon
1998; Delgado, Hopkins, and Kelly 1998). Studies of countries
across Africa and Asia also generally show that consumption
309
310

Table 17.10 Estimated Economic Impacts of Subsidy Programs in Malawi

Economic impact 2005/06 2006/07 2007/08 2008/09
Domestic maize prices ($/mt) May–Oct Nov 2006– May–Oct Nov 2007– May–Oct Nov 2008– May–Oct Nov 2009–
2006 Apr 2007 2007 Apr 2008 2008 Apr 2009 2009 Apr 2010
139.5 142.7 119.3 243.8 336.6 434.6 271.4 307.1
Maize price in benefit-cost and
fiscal efficiency analysis ($/mt) 143 154 250 280
Fertilizer price in analysis ($/mt) 393 490 590 1,250
Benefit-cost ratio, high response 1.38 1.30 1.90 1.08
Benefit-cost ratio, moderate
response 1.12 1.06 1.54 0.90
Benefit-cost ratio, low response 0.86 0.81 1.18 0.72
Fiscal efficiency, high response 0.76 0.44 1.13 0.09
Fiscal efficiency, moderate response 0.24 0.09 0.68 negative
Fiscal efficiency, low response negative negative 0.23 negative
2005 2006 2007 2008
Poverty incidence 50% 45% 40% 40%
Meals per day 2.0 2.2 2.3 2.3

Sources: SOAS et al. 2008; Dorward, Chirwa, and Slater 2010b; Food and Nutrition Security Program 2008; NSO 2006, 2009.
Note: The benefit-cost ratio and fiscal efficiencies were calculated with high, medium, and low fertilizer responses of 18, 15, and 12 kilograms of grain per kilogram of fertilizer, respectively. The
benefit-cost ratio is calculated as the gross incremental benefits divided by the gross incremental cost, valued at social prices. Fiscal efficiency is calculated as the net economic benefit divided
by fiscal cost.
Figure 17.6 Tracing Direct and Indirect Subsidy Impacts
Rural
households
Poorer
households
Resale
Input Incremental
subsidy use
Displacement
use
less poor
households
Y1 Increased
wages
Rural
economy
Source: SOAS et al 2008.
Note: Solid lines indicate positive impacts for poorer, food-deficit producers and sellers of labor; dashed lines indicate negative impacts for the less poor,
maize surplus producers, and buyers of labor.
linkages are more important than production linkages,
accounting for between 50 percent (in Senegal) and 98 per-
cent (in Zambia) of overall multipliers calculated (Delgado,
Hopkins, and Kelly 1998). These studies do not, however,
explicitly consider the further effects of transfers and
growth on seasonal capital constraints, or the knock-on
effects of relaxing these constraints.
The extent of these effects and of direct and indirect
increases in wages and benefits to poorer households as a
result of implementation of the subsidy program is, however,
an empirical question, as is the extent to which productivity
and welfare benefits are carried forward from one year to
the next. Focus group discussions in 2007, reporting on the
effects of the 2005/06 subsidy, clearly (and independently)
articulated processes of easing of seasonal cash constraints
in the hungry gap and tightening of labor markets, with
higher wages and low maize prices throughout the season
(SOAS et al. 2008). Similar focus group discussions con-
ducted in 2009 reported income benefits from higher wages
and increased maize availability, though they were also con-
cerned about the adverse effects of high maize prices—and
the high maize prices experienced after 2006/07 would be
expected to undermine these processes.
Livelihood and rural economy models have also been
used to investigate and describe this process in 2006/07 (see
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Farm/non-farm investment
Y1 Increased Y2 Increased
real incomes real incomes
Y1 Increased Y2 Increased
production production
Y2 Reduced Y2 Increased
maize prices wages
Input service demand Farm/nonfarm
and investment demand and investment
SOAS et al. 2008) and for 2008/09 (Dorward 2010). These
studies provide results consistent with those of the focus
group discussion reports, finding that the subsidy con-
tributes to higher wages but had lower wage effects in
2008/09 and that subsidy benefits have become more con-
centrated among beneficiaries (because the value of direct
production benefits to subsidy recipients is increased with
higher maize prices, but the benefits of lower maize prices
and higher wages to poor nonbeneficiaries are absent or
reduced). Fragmented information on wage rates over the
period 2005/06 to 2008/09 suggests that wage rate increases
over the period were higher than maize price increases over
the same period, representing real increase in wage rates.
Further empirical work is required to investigate the
scale of the indirect benefits through wage effects, because
they depend heavily upon the scale of growth multipliers
(consumption multipliers of slightly more than 1.5 embed-
ded in the livelihood and rural economy models are
derived from historical expenditure patterns in Malawi,
which are augmented by savings multipliers in the model).
Without these multipliers, wage impacts would be very
low—analysis of 2006/07 and 2008/09 household survey
data suggests that agricultural wage labor constitutes less
than 20 percent of household income even among the
poorest 20 percent of rural households in Malawi, although
311
 observed rises in agricultural wages suggest that unskilled,
nonagricultural wages have also risen, unless these two labor
markets are completely separated, an unlikely scenario.
The overall increase in real wages over the period is sup-
ported by anecdotal reports of rising rural wages and is con-
sistent with the limited political and social impact of and
response to high nominal maize prices in early 2009. It is
also consistent with reports of falling poverty rates and
decreased wasting among children under the age of five. The
subsidy program is not the only contributor to these
improvements. Over the same period, there were good rains,
a marked improvement in macroeconomic management,
and relatively high tobacco prices. These factors will have
made direct contributions to economic growth, and also
facilitated the implementation and impact of the subsidy
program (the tobacco prices and macroeconomic manage-
ment contribute to the availability of foreign exchange for
fertilizer imports, reduce the crowding-out effects of the
program, and stimulate other complementary parts of the
economy; good rains promote good yield responses to
fertilizers).
While attribution of these changes to the program is dif-
ficult, there are reasons to believe that the stimulus to maize
production from the subsidy and the good rains has been a
critical element in rural economic growth and poverty
reduction, as compared with the effects of good tobacco
prices alone. Limited evidence from elsewhere in Africa and
the world suggests that even labor-demanding smallholder
cash crops can drive pro-poor growth in such a broad way.
Further empirical work is also needed to establish the extent
to which productivity and welfare gains in one year are car-
ried forward to subsequent years, the conditions under
which these gains can be maximized, and the implications
for subsidy program design and implementation and ques-
tions about graduation.
CONCLUSIONS AND LESSONS
This chapter suggests that the Malawi agricultural input sub-
sidy program has achieved substantial benefits and successes,
although these are more nuanced than some press reports on
the program suggest. The measured producer benefits of the
program relative to its costs have been relatively modest;
however, our benefit-cost analysis does not capture all the
benefits of the program, nor all of its costs. Moreover, there
is scope for considerable improvement in the program’s
effectiveness and efficiency, although there are also practical
and political difficulties regarding the implementation of
some of these and questions about their effects.
312 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Other African countries considering the introduction of
agricultural input subsidies can learn important lessons
from Malawi’s experience. The program is a bold, large-
scale initiative that has achieved substantial increases in
maize production. The implementation of such a program
represents a very considerable logistical achievement, and
the government is to be commended for this and for its con-
tinuing and often imaginative attempts to improve the
program. Nevertheless, higher maize prices and calculation
of the agronomic yield effects of incremental fertilizer and
seed suggest the production increases resulting from the
program are not as large as might appear from the post-
2005/06 national maize production increases reported by
MoAFS. The benefits of the program also have to be
weighed against its very considerable costs (with an average
of over 9 percent of the national budget going to the subsidy
program since 2005/06) and the loss of benefits from alter-
native investment of these funds. While the food security
and growth benefits of the program have been partially
undermined by high maize prices, there have still been sig-
nificant improvements in productivity and welfare from
2005/06 associated with greater maize availability and with
increases in real wages.
Ongoing implementation challenges that the govern-
ment is working on include controlling costs, timing of
input deliveries, effective targeting of subsidized inputs,
reducing diversion and fraud, improving agronomic and
market returns with complementary investments (for
example, in extension, research, organic soil fertility
improvement, and roads), and using the subsidy program to
extend private sector input delivery systems. Success in
addressing these challenges will lead to new challenges
because increasing success will lead to changes in the need
for such a large-scale program with these objectives—and
continuing political, strategic, economic, technical, and
logistical system innovations will be needed to respond to
these changes.
Malawi’s experience with its large-scale agricultural
input subsidy program offers a number of important les-
sons to other countries in Africa considering the introduc-
tion of agricultural input subsidies.
First, any growth and development strategy that involves
agricultural input subsidies must be rooted in the opportu-
nities for and constraints to growth and development facing
a country and particular groups within it. This chapter has
set out in some detail specific difficulties that constrain
broad-based growth in Malawi, highlighting the reliance on
low-productivity maize and the difficulties and limited
options faced by very large numbers of poor, food-deficit
farmers, and indeed by the Malawian economy as a whole,
in breaking out of the low-maize-productivity and poverty
traps.
This low-productivity trap arises as a result of severe sea-
sonal credit constraints affecting very large numbers of
poor, food-deficit farming families, together with thin and
high-risk, high-margin input and maize markets. Malawi’s
key achievements with its subsidy program have been the
ability to raise land and labor productivity and improve
food security for large numbers of poor households by
relieving both profitability and affordability constraints on
the use of inputs needed to increase staple crop productiv-
ity, leading to some combination of increased real wages
and reduced food prices. The Malawian model thus applies
to other countries only if there are large numbers of people
facing similar staple-food-productivity constraints along-
side increased input use constrained by thin input markets,
poorly developed input supply systems, and widespread
profitability and affordability problems.
Malawi’s experience also shows that, in the right context,
large-scale agricultural input subsidy programs have the
potential to yield substantial benefits to people and their
governments with good design and implementation. The
chapter has also shown the very substantial costs and
resources required for such programs, and the difficulties
and challenges that must be overcome for effective, efficient,
and sustainable delivery of program benefits. Several issues
from Malawi’s experience are relevant to other countries
considering similar subsidy programs:
■ Focus: subsidies should be provided for inputs whose use
for important staple crops is constrained by affordability
difficulties despite high potential responses to input use.
■ Consumer gains: strong emphasis should be put on wider
contributions to economic growth and poverty reduc-
tion through consumer as well as producer gains.
■ Scale: the subsidy should affect staple crop prices, labor
markets, or both, requiring sufficient local or national
scale to affect markets, but strict limits on scale and the
control of costs are needed to limit displacement of exist-
ing purchases, crowding out of critical complementary
investments, and adverse macroeconomic impacts.
■ Logistical systems face major challenges in delivering
timely, targeted subsidies to large numbers of widely dis-
persed farmers, and the establishment of such logistical
systems requires time and major investments.
■ Performance monitoring, information, and auditing systems
are needed to develop trust, control fraud, and promote
efficiency and effectiveness. Debates on crop production
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
estimates and the number of farm families in Malawi
also demonstrate the importance of reliable information
for issues beyond specific matters related to the imple-
mentation of the subsidy program and assessments of its
impacts.
■ Effective targeting and rationing systems are needed to
limit scale and increase subsidy impacts on productivity,
but different (geographical or household) approaches
face different costs and difficulties (Dorward 2009), and
in some situations strict rationing of universal provision
may be a practicable alternative.
■ Entitlement systems are needed for targeting and
rationing, and these need to be robust against inevitable
counterfeiting and diversion.
■ Input supply system development requires close attention
to the complementary and changing roles and interests
of different public sector and commercial stakeholders,
but improved farmer access to input services should be a
major objective and outcome of agricultural input sub-
sidy programs.
■ Complementary policies and investments: if a subsidy pro-
gram is seen as part of a broad, long-term strategy for
poverty reduction and economic development, then
investments in complementary activities must be made
in areas such as extension, research, organic soil fertility
improvement, health, education, markets, transport and
communication infrastructure, and services. Considera-
tion of the different roles of these complementary invest-
ments should also guide decisions on the nature, scale,
and implementation of the input subsidy, as well as of
other investments, in order to achieve positive interac-
tions among investments.
■ Macroeconomic management to promote favorable
growth conditions and provide budgetary resources
needed for such a program is also important.
■ Political commitment is required for sustained mobiliza-
tion of program resources, but there may also be poten-
tial conflicts between the need for political support on
the one hand and targeting, rationing, cost control, and
performance monitoring needed for efficient and sus-
tainable implementation on the other.
■ Sustainability of program implementation should be
addressed by attention to cost control, scale, and logistical
and performance monitoring and audit systems. There
is also need for investigation of sustainability of impacts,
with examination of the extent to which productivity and
welfare gains carry forward from one year to the next and
the implications of this analysis for program design
and implementation and for questions about graduation.
313
 An innate dilemma in the design and implementation of
large-scale subsidy programs is that such programs require
both stability and flexibility, with innovation. Stability is
needed to provide stakeholders with confidence and security
that will justify long-term financial and other investments
associated with the program’s implementation. Stability can
be undermined, however, by the need for flexibility to adjust
to changing conditions (for example in the weather, in
international and national markets and economies, or in
politics), and some of these changes may be anticipated or
unanticipated results of the program. Alongside flexibility is
the need for innovation (in technology, systems, and prices)
to take advantage of learning and change during program
implementation. Although flexibility and innovation can
undermine stability, lack of flexibility and innovation may
also undermine stability if conditions, such as increasing
incidence of fraud, make the system unsustainable and inef-
fective in its initial form. To achieve mutually supportive
stability, flexibility, and innovation, trust and stable princi-
ples must govern both the long-term objectives of and rela-
tions between different stakeholders on one hand and the
processes for successful learning, flexibility, and innovation
on the other.
NOTES
1. This section draws heavily on material from SOAS
(2008).
2. Reasons for the high dependency on maize as opposed
to other food crops include dietary preferences, different
crops’ relative calorific yields per hectare in different agro-
ecologies, farmers’ familiarity with the crop, and long-
standing strong government policies aimed at promoting
maize production and input and crop marketing subsidies
focused on maize.
3. Understanding of the nature, causes, and relative impor-
tance of these problems varies (indeed, elements of the
analysis presented here are not universally accepted, nor is
this paper a comprehensive account of the complex issues
involved).
4. The postharvest value-to-cost ratio has generally been
less than 2, widely considered to be the minimum required
to make fertilizer use profitable in moderately but not
highly risky situations (Morris et al. 2007). Even for higher
preharvest maize prices, the ratio typically has been around
or below 2 (SOAS 2008), depending on the yield response
achieved. With high yield responses, the value-to-cost ratio
has been above 2 in some years (Maize Productivity Task
Force 1997).
5. Table 17.1 shows, for example, that the median for
maize stocks running out each year is between four and six
314 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
months after harvest, and that in 2003/04 one bag of fertil-
izer represented approximately 10 percent of rural house-
holds’ median per capita annual expenditure (and more
than 20 percent of median per capita expenditure of the
lowest expenditure quintile).
6. A ratio of 15 was used in the calculations cited above
based on what are believed to represent mean grain-to-
nitrogen response rates to fertilizer application on farmers’
fields. Improved management and uptake of hybrid seed
provide the potential for higher ratios (in the range of 22 to
28 kilograms of maize per kilogram of fertilizer applied).
Hence, improved farm management practices have the
potential to make fertilizer use on maize profitable even
without subsidy, although the affordability constraint
remains.
7. This section draws heavily on Dorward (2009).
8. This analysis applies only to subsidies implemented on a
scale large enough to affect output prices. Small-scale subsi-
dies that do not significantly affect product prices are equiv-
alent to highly elastic product demand: subsidy benefits are
largely captured by suppliers and producers, and dead-
weight costs depend upon the elasticity of supply.
9. This is not a problem in situations in which the
providers of land and labor are poor. Indeed, subsidies can
promote pro-poor growth in such situations.
10. For this, complementary constraints to effective use of
inputs by targeted beneficiaries must be addressed.
11. The terms “coupon” and “voucher” are used inter-
changeably in wider discussions of the Malawi Agricultural
Input Subsidy program and in this paper.
12. Efficient and stable output markets may require less
government intervention in direct market operations but
more focused investment in market insurance and in facili-
tating infrastructure and institutions.
13. This section draws heavily on SOAS (2008) and on
Dorward and Chirwa 2009a.
14. No information is available on seed sales.
15. A standard maize subsidy pack in 2005/06 consisted of
one voucher for a 50 kilogram bag of 23:21:0+4S, one
voucher for a 50 kilogram bag of 23:21:0+4S, and one
voucher for improved maize seed. A standard tobacco sub-
sidy pack consisted of one voucher each for a 50 kilogram
bag of calcium ammonium nitrate (CAN) and a 50 kilo-
gram bag of D Compound. The fertilizer vouchers were
redeemable for MK 950. The seed voucher required no top-
up and could be used to purchase 2 kilograms of hybrid
seed or 3 to 4 kilograms of OPV seed, depending on the
price set by the seed supply company.
16. These variations in targeting stem from vagueness in the
definition of target beneficiaries in the guidelines and dif-
ferences in the way communities dealt with problems of
shortages. Variations meant that those that were targeting
beneficiaries placed different emphasis on different criteria
and processes.
17. Displacement for 2007/08 is assumed to be similar to that
for 2006/07 (higher subsidy sales may increase displacement,
but greater farmer familiarity and higher fertilizer prices
would be expected to reduce displacement). A lower displace-
ment is assumed for 2008/09 as a result of much higher fer-
tilizer prices and earlier (separate) beneficiary registration.
18. Low average prices in some years in the early 1990s were
brought about by large-scale imports, and at the turn of the
century by the starter pack subsidy program (with good
weather).
19. Although the annual average price in 2008/09 was very
high, peak prices were equivalent to those in 2001/02 and
2005/06 because prices rose much earlier in the season in
2008/09 but then flattened out.
20. The latter may be higher if there are significant exports
or commercial carryover stocks at the end of the season.
21. These include inherent and difficult-to-quantify biases
in different methods of collecting yield data, multicollinear-
ity between input use and other management variables, high
variability in smallholder agriculture, and variation of
response rates with average rates of subsidized fertilizer use.
22. This calculation considers only maize production.
Provision of the fertilizer subsidy for tobacco leads to a
much higher displacement than for maize; thus, assuming
that all incremental fertilizer is used on maize is not
unreasonable. Fertilizer impacts on crops mixed with
maize are ignored here.
23. It is important to note that increased hybrid seed subsi-
dies may not be required if farmers are able and willing to
purchase unsubsidized seed—improved access to seed,
increased distributors, and effective extension may be more
effective in increasing hybrid seed use.
24. For 2007/08, however, low crop output resulting from
local droughts and floods affected total production, not just
incremental production from the subsidy.
REFERENCES
Ariga, J., and T. Jayne. 2009. “Private Sector Responses to
Public Investments and Policy Reforms. The Case of Fer-
tilizer and Maize Market Development in Kenya.” Dis-
cussion Paper 00921. International Food Policy Research
Institute (IFPRI), Washington, DC.
Ariga, J., T. Jayne, B. Kibaara, and J. K. Nyoro. 2008. “Trends
and Patterns in Fertilizer Use by Smallholder Farmers in
Kenya, 1997–2007.” KeWP 28, Tegemeo Institute, Nairobi.
Chapoto, A., and T. S. Jayne. 2010. “Maize Price Instability in
Eastern and Southern Africa: The Impact of Trade Barriers
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
and Market Interventions.” Prepared for the COMESA
policy seminar “Variation in Staple Food Prices: Causes,
Consequence, and Policy Options.” Maputo, Mozam-
bique, 25-26 January 2010 under the Comesa-MSU-
IFPRI African Agricultural Markets Project.
Chirwa, E., M. Matita, and A. Dorward. 2010. “Factors
Influencing Access to Agricultural Input Subsidy
Coupons in Malawi.” Future Agricultures Consortium,
Brighton, U.K.
Crawford, E. W., T. S. Jayne, and V. A. Kelly. 2006. “Alterna-
tive Approaches for Promoting Fertilizer Use in Africa.”
Agriculture and Rural Development Discussion Paper
22, World Bank, Washington, DC.
Delgado, L. C., J. Hopkins, and V. A. Kelly. 1998. “Agricul-
tural Growth Linkages in Sub-Saharan Africa.” IFPRI
Research Report. IFPRI, Washington, DC.
Djurfeldt, G., H. Holmen, M. Jirstrom, and R. Larsson, eds.
2005. The African Food Crisis: Lessons from the Asian
Green Revolution. Wallingford, U.K.: CABI Publishing.
Dorward, A. R. 2006. “Markets and Pro-poor Agricultural
Growth: Insights from Livelihood and Informal Rural
Economy Models in Malawi.” Agricultural Economics 35
(2): 157–69.
———. 2009. “Rethinking Agricultural Input Subsidy Pro-
grams in Developing Countries.” Non-Distorting Farm
Support to Enhance Global Good Production, ed. A.
Elbehri and A. Sarris, pp. 311-74. Rome: Food and Agri-
culture Organization. Also available at http://www.fao
.org/es/esc/common/ecg/586/en/Dorward_FAO_Subsidy_
Paper_FINAL.pdf.
———. 2010. “Informal Rural Economy Modelling of Sub-
sidy Impacts.” Unpublished paper, SOAS (School of Ori-
ental and African Studies), London.
Dorward, A. R., and E. Chirwa. 2009a. “The Agricultural
Input Subsidy Program 2005 to 2008: Achievements and
Challenges.” School of Oriental and African Studies,
London.
———. 2009b. “The Malawi Agricultural Input Subsidy
Programme: 2009 Fieldwork Preliminary Report.”
School of Oriental and African Studies, London.
———. 2010. “Evaluation of the 2008/09 Agricultural Input
Subsidy Malawi: Maize Production Impacts.” School of
Oriental and African Studies, London.
Dorward, A., E. Chirwa, and R. Slater. 2010a. “Evaluation of
the 2008/09 Agricultural Input Subsidy Malawi: Prelimi-
nary Report on Program Impact.” School of Oriental and
African Studies, London.
———. 2010b. “Evaluation of the 2008/09 Agricultural
Input Subsidy Malawi: Report on Program Implemen-
tation.” School of Oriental and African Studies,
London.
315
 Dorward, A. R., J. G. Kydd, J. A. Morrison, and I. Urey. 2004.
“A Policy Agenda for Pro-Poor Agricultural Growth.”
World Development 32 (1): 73–89.
Economist Intelligence Unit. 2008. “Lifting African and
Asian Farmers Out of Poverty: Assessing the Investment
Needs.” Custom research project for the Bill and Melinda
Gates Foundation, New York.
Ellis, F. 1992. Agricultural Policies in Developing Countries.
Cambridge, U.K.: Cambridge University Press.
Fan, S., A. Gulati, and S. Thorat. 2007. “Investment, Subsi-
dies, and Pro-Poor Growth in Rural India.” Discussion
Paper 716, IFPRI, Washington, DC.
Fan, S., S. Thorat, and N. Rao. 2004. “Investment, Subsidies,
and Pro-Poor Growth in Rural India: Institutions
and Economic Policies for Pro-poor Agricultural
Growth.” IFPRI discussion paper, IFPRI, Washington,
DC.
FAO (Food and Agriculture Organization). 2009. FAO Agri-
cultural production data. Rome.
FEWS NET. 2009. “Malawi Food Security Outlook, October
2009–March 2010.” Lilongwe.
Food and Nutrition Security Program. 2008. “Monitoring
and Evaluation, 6th Report,” (draft). Lilongwe.
Govereh, J., J. Shawa, E. Malawo, and T. S. Jayne. 2006. “Rais-
ing the Productivity of Public Investments in Zambia’s
Agricultural Sector.” Working Paper 20, Food Security
Research Project, Lusaka.
Harrigan, J. 2003. “U-Turns and Full Circles: Two Decades
of Agricultural Reform in Malawi 1981–2000.” World
Development 31 (5): 847–63.
Hazell, P., and B. Hojjati. 1995. “Farm/Non-Farm Growth
Linkages in Zambia.” Journal of African Economies 4 (3):
406–35.
Hazell, P., and M. Rosegrant. 2000. Rural Asia: Beyond
the Green Revolution. Manila: Asian Development
Bank.
Jayne, T. S., J. Mangisoni, N. Sitko, and J. Ricker-Gilbert.
2010. “Malawi’s Maize Marketing System.” Report com-
missioned by the World Bank and Ministry of Agriculture
of the Government of Malawi, Lilongwe.
Kelly, V. A., D. Boughton, and N. Lenski. 2010. “Malawi Agri-
cultural Inputs Subsidy Program: Evaluation of the
2007/08 and 2008/09 Input Supply Sector Analysis.”
Michigan State University, East Lansing, MI.
Levy, S., ed. 2005. Starter Packs: A Strategy to Fight Hunger
in Developing and Transition Countries? Lessons from the
Malawi Experience, 1998–2003. Wallingford, U.K.:
CABI.
Logistics Unit. 2009. “Final Report: Implementation of Agri-
cultural Inputs Subsidy Program 2008/09.” Lilongwe.
316 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Maize Productivity Task Force, Ministry of Agriculture of
Malawi. 1997. “Area-specific Recommendations for
Hybrid Maize Grown by Malawian Smallholders: A Man-
ual for Extension Field Assistants.” Lilongwe.
Mangisoni, J. 2010. “Defects and Storage Loss Analysis of
Maize Grain in Malawi.” Unpublished paper, Bunda Col-
lege, University of Malawi.
MoAFS (Ministry of Agriculture and Food Security of
Malawi). 2008. “The 2007/2008 Inputs Subsidy Program
Review Report: An Internal Review.” Department of
Agricultural Planning Services, Lilongwe.
Minot, N. 2009. “Staple Food Prices in Malawi.” Prepared
for the COMESA policy seminar “Variation in Staple
Food Prices: Causes, Consequences, and Policy Options.”
Maputo, 25–26 January 2010 under the African Agricul-
tural Marketing Project, IFPRI.
Morris, M., V. A. Kelly, R. Kopicki, and D. Byerlee. 2007.
“Fertilizer Use in African Agriculture.” World Bank,
Washington, DC.
Munthali, M. W. 2007. “Integrated Soil Fertility Manage-
ment Technologies: A Counteract to Existing Milestone
in Obtaining Achievable Economical Crop Yields in Cul-
tivated Lands of Poor Smallholder Farmers in Malawi.”
In Advances in Integrated Soil Fertility Management in
Sub-Saharan Africa, ed. A. Bationo, B. Waswa, J. Kihara,
and J. Kimetu, pp. 533–536. Dordrecht, The Netherlands:
Springer.
Nakhumwa, T. O. 2006. “Rapid Evaluation of the 2005 Fer-
tilizer Subsidy Program in Malawi.” Policy Paper No 10,
CISANET, Lilongwe.
NSO (National Statistical Office of Malawi). 2005. “Inte-
grated Household Survey 2004 –2005.” Zomba.
———. 2006. “Welfare Monitoring Survey 2005.” Zomba.
———. 2008. “Welfare Monitoring Survey 2008.” Zomba.
———. 2009a. “2008 Population and Housing Census: Pre-
liminary Report.” Zomba.
———. 2009b. “Welfare Monitoring Survey 2008.”
Zomba.
Reardon, T. 1998. Rural Non-farm Income in Developing
Countries: The State of Food and Agriculture 1998. Rome:
Food and Agriculture Organization.
Reserve Bank of Malawi. 2010. “Financial and Economic
Review 2009.” Lilongwe.
Ricker-Gilbert, J., and T. Jayne. 2010. “The Impact of Fertil-
izer Subsidies on Displacement and Total Fertilizer Use.”
Unpublished presentation, Lilongwe.
Ricker-Gilbert, J., T. S. Jayne, and E. Chirwa. 2011. “Subsidies
and Crowding Out: A Double-Hurdle Model of Fertilizer
Demand in Malawi.” American Journal of Agricultural
Economics 93 (February): 26–42.
SOAS (School of Oriental and African Studies), Wadonda
Consult, Overseas Development Institute, and Michigan
State University. 2008. “Evaluation of the 2006/07 Agri-
cultural Input Supply Program Malawi: Final Report.”
SOAS, London.
South African Revenue Service. 2009. Monthly maize quan-
tities exported to Malawi. SARS database. Pretoria.
Siamwalla, A., and A. Valdes. 1986. “Should Crop Insur-
ance Be Subsidized?” In Crop Insurance for Agricultural
Development: Issues and Experience, ed. P. Hazell, C.
Pomareda, and A. Valdes, pp. 117–25. Baltimore: IFPRI
and John Hopkins University Press.
CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09
Timmer, C. P. 1989. “Indonesia: Transition from Food
Importer to Food Exporter.” In Food Price Policy in Asia,
ed. T. Sicular, pp. 22–64. Ithaca, NY: Cornell University
Press.
World Bank. 1981. “Accelerated Development in Sub-Saharan
Africa: An Agenda for Action.” World Bank, Washington,
DC.
Xu, Z., W. A. Burke, T. S. Jayne, and J. Govereh. 2009. “Do
Input Subsidy Programs ‘Crowd In’ or ‘Crowd Out’ Com-
mercial Market Development? Modeling Fertilizer
Demand in a Two-Channel Marketing System.” Agricul-
tural Economics 40 (1).
317
 CHAPTER 18
MoneyMaker Pumps: Creating Wealth
in Sub-Saharan Africa
I. V. Sijali and M. G. Mwago
he first of the Millennium Development Goals
T
(MDGs) is to eradicate extreme poverty and
hunger, with the target of halving the proportion
of people whose income is less than $1 a day between 2000
and 2015. In Sub-Saharan Africa, rural poverty accounts
for 83 percent of total extreme poverty, with 85 percent
of the poor depending partly on agriculture for their
livelihoods (World Bank 2000). Considering that agricul-
ture can have the greatest impact on poverty and food
security if the benefits of its development are reaped by
the poor, priority should be given to the development of
small-scale irrigation (Kidane, Maetz, and Dardel 2006).
Agricultural growth is therefore a key to the realization of
the first MDG.
Stimulating agricultural growth is critical to reducing
poverty in Africa. Commercial agriculture, potentially a
powerful driver of agricultural growth, can develop along a
number of pathways. But agricultural productivity in Sub-
Saharan Africa is the lowest in the world, with per capita
output only 56 percent of the world average (FAO 2005),
and agricultural output in Sub-Saharan Africa has not kept
pace with population increase (Msangi and Rosegrant
2005). More than 80 percent of output growth in Sub-Saha-
ran Africa since 1980 has come from expansion of the
cropped area, compared with less than 20 percent for all
other regions. Food self-sufficiency among Sub-Saharan
Africans declined from 97 percent in the mid-1960s to
82 percent in 1997–99, without a substantial increase in
household incomes that would allow people to afford food
purchases. Agricultural families have remained locked in a
low-input, low-income system, with low and stagnating
agricultural yields.
Agricultural intensification has become the main avenue
of growth in crop production, with agricultural water man-
agement (water harvesting and irrigation) the focus of this
intensification in Sub-Saharan Africa. More reliable access
to water for irrigation would stabilize yield, reduce the risk
of harvest failure, and increase productivity and agricul-
tural growth. Successful investment in agricultural water
management technologies therefore presents an important
opportunity for poverty reduction and economic growth.
However, major constraints to using irrigation—including
lack of surface storage structures, appropriate irrigation
technologies, capacity to exploit underground water, com-
munity skills for irrigation, information on appropriate
irrigation technologies, and investment in irrigation—
must be addressed.
In some parts of Sub-Saharan Africa, investment in
irrigation development at the community level is already
bearing fruits. In Tanzania the Participatory Irrigation
Development Project has had striking impacts, increasing
farm income by 86 percent and enabling irrigator house-
holds to enjoy better-quality housing, acquire agricultural
and household assets, access health services, and finance
children’s education (IFAD 2005). In four representative
subproject areas, ownership of oxcarts and cattle
increased considerably, the total number of grinding mills
increased from 2 to 12, and the number of shops
319
 increased from 2 to 74. Similarly, in Zimbabwe, irrigator
households in the European Union-funded Maunganidze
Irrigation Scheme reported an increase in income of more
than 200 percent and turned a food deficit into a surplus.
Investment in new housing and in water and sanitation
was the most obvious sign of improved livelihoods, with a
number of modern two- and three-room houses with
ventilated pit latrines and in several cases protected water
wells. IFAD concluded that these impacts were the result
of investment in irrigation, because there were no other
sources of income in the area.
Studies have shown that small-scale irrigation offers
opportunities to improve livelihoods in Africa. However,
there has been little technology-driven jump in productivity
for the majority of resource-poor African farmers. IFAD
(2005) estimates that of the potentially irrigable 39.4 million
hectares in Africa, only 7.1 million hectares, or 18 percent
of the total, have been equipped for irrigation. Expansion
of irrigation has been slow. Over the past forty years, only
4 million hectares have become irrigated in Africa, by far
the smallest expansion of any world region. The area
developed for irrigation as a fraction of the irrigation
potential for African countries is shown in map 18.1. A
majority of African countries have exploited less than
50 percent of their potentially irrigable land, with 10
countries at less than 10 percent, 8 countries at 10 to 25
percent, and 6 countries at 26 to 50 percent. Eleven coun-
tries have exploited 50–75 percent, while only 2 countries
have exploited more than 75 percent of their potential
(FAO 2005).
The lack of exploitation of irrigation potential in Sub-
Saharan Africa can be attributed to lack of access to
appropriate and affordable irrigation technologies by the
majority of farmers in these countries. The cost per
hectare of developing land for irrigation in Sub-Saharan
Africa ranges from $2,000 to $4,000 for small-scale irriga-
tion and from $9,000 to $15,000 for large-scale irrigation,
making it out of reach for most families and communities
(Kidane, Maetz, and Dardel 2006). In India, the compara-
ble cost for large-scale irrigation ranges from $1,500 to
$2,000 per hectare.
In recent years, however, the development of appropri-
ate and affordable irrigation technologies for poor Sub-
Saharan Africa farmers and communities has been
remarkable. One of these technologies is the human-
powered irrigation pumps developed by KickStart Inter-
national, which have been used by dynamic farmers and
entrepreneurs to establish and run profitable small-scale
agribusinesses.
320 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA
KICKSTART INTERNATIONAL
KickStart International is a nonprofit social enterprise
organization founded in Kenya in 1991. It now operates in
Burkina Faso, Kenya, Mali, and Tanzania. KickStart’s mission
is to promote economic growth and employment creation in
Africa by developing and promoting technologies that entre-
preneurs can use to establish and run profitable small-scale
businesses. The entrepreneurs raise small amounts of capital
($100–$1,000) to start a new enterprise, and KickStart helps
them to identify viable business opportunities and obtain the
appropriate technologies required.
KickStart addresses poverty alleviation through devel-
opment of appropriate technologies and innovations to
enhance production, value addition, and income genera-
tion. These include, among others, irrigation pumps, an
oil press, and a brick press. KickStart’s irrigation pumps
enable smallholder farmers to enhance productivity and
improve household incomes and thus contribute to sus-
tainable poverty reduction. KickStart has developed several
human-powered pumps. These pumps have the distinct
advantage of being operable without fossil fuels or electric-
ity, which is not always available in remote areas.
The business model
KickStart’s business model is a fivefold strategy:
■ Market research for business opportunities. KickStart
identifies business opportunities through market
research and feasibility studies that are aimed at assist-
ing the rural poor.
■ Research and development. KickStart designs tools to end
poverty that can be used by small-scale entrepreneurs to
start their own profitable businesses.
■ Private manufacturing. After developing the tools, private
manufacturers are contracted to produce the products in
a way that ensures high-quality standards and optimal
pricing for customers. KickStart’s manufacturers are
located in China, Kenya, and Tanzania, with the bulk of
the products originating from China and, to a lesser
extent, from Kenya. As of 2010, the manufacturer in
Tanzania is capable of supplying only a portion of the
Tanzania demand.
■ Marketing and distribution. KickStart’s products are
introduced to the market through its own marketing and
promotional activities, and the private sector supply
chain is used to distribute the products locally. KickStart
has more than 160 distributors in Kenya, more than
150 in Tanzania, more than 80 in Mali, and a few in
Map 18.1 Irrigated Areas as a Percentage of Potentially Irrigable Areas

IBRD 38554
MAY 2011
30° 0° 30° 60°

Mediterranean
TUNISIA Sea
MOROCCO
30° Canary Is. 30°
(Sp)
ALGERIA ARAB
LIBYA
FORMER REP. OF
SPANISH EGYPT
SAHARA

Red
CAPE MAURITANIA

Sea
VERDE
MALI NIGER
ERITREA
SENEGAL CHAD
THE GAMBIA BURKINA SUDAN
DJIBOUTI
GUINEA-BISSAU GUINEA FASO BENIN
NIGERIA ETHIOPIA
GHANA
TOGO

SIERRA LEONE CÔTE

CENTRAL
LIBERIA D’IVOIRE AFRICAN REPUBLIC
CAMEROON
EQUATORIAL GUINEA SOMALIA
UGANDA
SÃO TOMÉ AND PRÍNCIPE KENYA
0° GABON CONGO DEMOCRATIC 0°
ATLANTIC REPUBLIC RWANDA
BURUNDI INDIAN
OF CONGO
OCEAN Cabinda
TANZANIA OCEAN
(ANGOLA)
SEYCHELLES
COMOROS
WATER MANAGED AREAS AS % ANGOLA MALAWI
ZAMBIA Mayotte
OF IRRIGATION POTENTIALS: (Fr)
76% - 100%
ZIMBABWE MOZAMBIQUE MADAGASCAR
51% - 75% NAMIBIA BOTSWANA MAURITIUS
26% - 50%
10% - 25% SWAZILAND
< 10% SOUTH LESOTHO
30° 30°
AFRICA
INTERNATIONAL BOUNDARIES

0 500 1,000 1,500 KILOMETERS

Source: Aquastat 2005 0 500 1,000 MILES

30° 0° 30° 60°

Source: FAO 2005.

Burkina Faso. In countries in which KickStart does not
have a marketing program, distributors have been hired.
KickStart also liaises with local and international non-
governmental organizations (NGOs) and with United
Nations agencies to serve those organizations’ beneficiar-
ies and programs.
■ Impact assessment. KickStart conducts socioeconomic
impact assessments on the incremental benefits resulting
from its products. Information gleaned from these
assessments is used to make improvements to the tech-
nology and to generate ideas for future innovation.
Very often, KickStart is asked whether it would be better
or more effective to simply give the pumps away. In both
cases, the answer cases is no. The KickStart model of selling
pumps through a profitable supply chain has been shown
to be both profitable and sustainable in many African
countries, creating employment opportunities and new

CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 321

 sources of income. Analysis has shown that the KickStart other than Kenya and Tanzania, the revenue data show an
model increases the livelihood of farmers and their families increase of 440 percent for the MoneyMaker pump.
more cost effectively than giving the pumps away would.
Annex 18.1 provides further details on this topic.
Impacts of pumps at the manufacturer and
KickStart uses a tipping-point concept to mark the
retailer levels
beginning of sustainable technology infusion. The tipping
point is where the volume of sales of the technology dra- Profits have been generated by the sales and manufactur-
matically rises or increases, decreasing the marketing ing of the MoneyMaker range of pumps at the manufac-
costs to a minimum, as illustrated in figure 18.1. KickStart turer and retailer levels, as shown in table 18.2. Skilled
estimates that the tipping point is achieved when 15 to
20 percent of the market potential in terms of sales is
Figure 18.1 The Tipping Point Concept
achieved. In Kenya, for example, KickStart aims to surpass
the tipping point by 2014.

Impacts of MoneyMaker pumps at the farm level Sales

The MoneyMaker pump technology has had a significant

impact on revenue generation and wages at the farm, dis-

Money
tributor, and manufacturer levels. Table 18.1 summarizes Production and
Sustainable
survey data of the income and wage impacts of KickStart’s marketing costs
business
different types of pumps at the farm level. As shown in the
table, household income increased by 191–200 percent in
Kenya for all categories of pumps except for the Money- Tipping
point
Maker Hand Pump, which showed no change. In Tanzania,
income doubled for households that used the MoneyMaker
Time
Hip Pump and tripled for households that used the Super
MoneyMaker and MoneyMaker Plus pumps. For countries Source: Authors.

Table 18.1 Farm-Level Household Revenues and Wages per Pump, July 1991–November 2009
(dollars)
Annual
income Annual Percent Wages over
without income increase Annual lifespan of
Pump pump with pump in income wages pumpa
Kenya
MoneyMaker 628 1,885 200 126 379
Super MoneyMaker and Super MoneyMaker Plus 1,148 3,443 200 71 214
MoneyMaker Plus 248 744 200 21 63
MoneyMaker Hand 93 279 200 4 12
MoneyMaker Hip 832 2,423 191 60 174
Tanzania
Super MoneyMaker and MoneyMaker Plus 493 1,478 200 41 123
MoneyMaker Hip 826 1,651 100 3 6
Other countriesb
MoneyMaker pump 183 995 444 18 54
Super MoneyMaker and Super MoneyMaker Plus 345 1,034 200 29 86
MoneyMaker Hand 38 — — 2 6
MoneyMaker Hip 578 1,156 100 3 9
Source: KickStart International.
a. Wages over lifespan of pump = annual wages x lifespan of pump.
b. Burundi, Democratic Republic of Congo, Malawi, Mali/Burkina Faso, Mozambique, Rwanda, Sudan, Uganda and Zambia.
— = Data not available.

322 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA

 Table 18.2 Manufacturer and Retailer Profits and Manufacturer Skilled Jobs, July 1991
to November 2009
Manufacturing Retailing
Technology Pretax profit ($) Skilled man days Pretax profit ($)
Kenya
MoneyMaker 310,694 1,738 20,918
Super MoneyMaker 61,378 10,391 155,882
Super MoneyMaker Plus 75,604 12,799 215,000
MoneyMaker Plus 33,999 3,837 58,284
MoneyMaker Hand Pump 27,712 1,564 6,455
MoneyMaker Hip Pump 33,197 2,997 37,844
Total 542,584 33,325 494,382
Tanzania
MoneyMaker 26,874 219 2,239
Super MoneyMaker 38,248 2,878 35,306
Super MoneyMaker Plus 269,024 20,241 292,034
MoneyMaker Plus 985 132 985
Money maker Hand Pump 925 108 1,013
MoneyMaker Hip Pump a a 17,332
Total 311,870 23,578 348,910
Source: KickStart.
a. Not manufactured in the country.

manpower jobs have also been created in the manufactur-
ing of the pumps. Between July 1991 and November 2009,
retailers’ profits from MoneyMaker pumps totalled
$494,382 in Kenya and $348,910 in Tanzania. At the man-
ufacturer level, profits over the same period amounted to
$542,584 in Kenya and $311,870 in Tanzania. The manu-
facturing of the pumps involved more than 33,000 and
nearly 24,000 skilled man days in Kenya and Tanzania,
respectively.
MONEYMAKER PUMPS PRODUCT LINE
Criteria for pump creation
KickStart is guided by a unique set of criteria in designing
its products: income generation, return on investment,
affordability, energy efficiency, portability, ease of use and
installation, strength and durability, design for manufac-
turing, and cultural acceptability. These criteria are
defined as follows:
■ Income generation. Every pump model developed and
marketed must have a business model that clearly pre-
dicts profitability and can be supplied to entrepreneurs.
■ Return on investment. A purchaser of a MoneyMaker
pump should be able to fully recoup his or her invest-
ment within six months.
■ Affordability. Because the target purchasers of the
pumps are some of the world’s poorest people, designs
and manufacturing must ensure that retail prices are
affordable, ideally less than $150.
■ Energy efficiency. All pumps are human powered and
thus must be extremely efficient at converting human
power to mechanical power
■ Ergonomics and safety. The pumps must be safe to use for
long periods of time without stress or injury.
■ Portability. Pumps must be small and light enough to
carry home from the point of purchase on foot, by bike,
or by minibus.
■ Ease of installation and use. All pumps must be easy to set
up and use, without additional training or tools, not even
a hammer or screwdriver.
■ Strength and durability. The pumps are designed and
built to withstand heavy use. The pumps carry a one-year
guarantee upon purchase.
■ Design for manufacturing. For manufacturing to be truly
effective, pumps must be produced in large quantities,
but in the developing world manufacturing capacity is
limited. The design of pumps must be such that the
pumps can be manufactured within the capacity of the
local manufacturing industry.
■ Cultural acceptability. The pumps must be adapted to the
culture in the countries where they are sold.
A team of engineers, designers, and technicians develop
and test prototype pumps to ensure their performance, cul-
tural acceptability, and durability. In addition to design and

CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 323

 development, KickStart creates marketing awareness and Figure 18.2 Design Diagram and Field Operation of the
assembles sales teams to sell the pumps. Importantly, Kick- Super MoneyMaker Pump
Start employs innovative marketing techniques that are
accessible to potential customers in remote villages without Super-MoneyMaker
fossil fuels and electricity, including community pumping pressure irrigation pump
competitions during market days and demonstrations at
retail shops and individual farms. Operation
Manual
Types of pumps

Original MoneyMaker Pump. KickStart’s debut pump,

the original MoneyMaker, was introduced in September Plastic Bearing
1996. This small, treadle-operated pump could pull water
Pressure
from as deep as 7 meters and be used to furrow irrigate up hose
pipe
to 0.8 hectares of land. The pumps demonstrate the poten-
Two rubber caps
tial poverty reduction effects of micro-irrigation: more than Two rubber valve (inside each cyclinder)
4,050 original MoneyMaker pumps were sold between 1996 (inside each cyclinder) Outlet pipe
and 1999, generating more than $3.9 million annually Rigid
Inlet pipe
function
among user households in East Africa. User feedback, how- pipe Outlet pipe
ever, indicated that farmers needed a pump that could lift Foot valve
(below water) Rubber valve
and push water through a hosepipe or sprinklers. In (inside)
response, KickStart introduced the new Super MoneyMaker Plastic pipe

pump in October 1998. The new suction and pressure Source: KickStart International.
pump superseded the original MoneyMaker, which was
taken off the market in February 1999.

Super MoneyMaker Pump. The Super MoneyMaker

pump is used to pump water from hand-dug wells, rivers, Figure 18.3 Design Diagram and Field Operation of
streams, lakes, and ponds. A twin-cylinder pump, it is ideal MoneyMaker Plus Pump
for sprinkler and drip irrigation, for filling overhead water
tanks, and for use with nozzles and sprays attached to the
end of the delivery hose. The pump can draw water from a
depth of 7 meters and pump it 14 meters above the
ground. It can be used to irrigate up to 0.8 hectares of
land. The design and operation of the Super MoneyMaker
pump is shown in figure 18.2. An improved, easier-to-use
design of the Super MoneyMaker pump, Super Money-
Maker Plus, was subsequently introduced.

MoneyMaker plus pump. Responding to demand for a

lower-cost pressure irrigation pump, KickStart designed
and launched the MoneyMaker Plus pump in July 2001.
This small, leg-operated pump has one piston and one
cylinder but can still pull water from 7 meters deep and
push it more than 14 meters above the ground. It can be
used to irrigate up to 0.4 hectares of land. The design and
operation of the MoneyMaker Plus pump is shown in
figure 18.3. Source: KickStart International.

324 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA

MoneyMaker Hip Pump. The MoneyMaker Hip Pump is
a unique pump developed to meet the needs of the very
poor people who cannot afford the twin-cylinder Super
MoneyMaker pump or the single-cylinder MoneyMaker
Plus. The MoneyMaker Hip Pump debuted in stores in
2006, however, sales and marketing efforts began in 2008.
The pump weighs only 4.5 kilograms, and can irrigate up to
0.4 hectares of land.
The design of the MoneyMaker Hip Pump includes a
super-efficient valve box and a simple pivot hinge. By
attaching a hand pump to a hinged platform, the mechan-
ics of the pump were changed to allow operators to use
their legs, body weight, and momentum, rather than the
few muscles of the upper back and shoulders. This makes
the pump more energy efficient, and it is capable of irri-
gating at least 0.4 hectares. Like the Super MoneyMaker,
the MoneyMaker Hip Pump can draw water from 7 meters
and pump up to 14 meters above the ground.
These pumps retail for $35 to $100. Between July 1991 and
November 1999, more than 137,000 MoneyMaker pumps of
all types were sold, with a vast majority of them purchased by
residents of African countries. Kenya, Tanzania, and Malawi
are the countries in which the greatest number of pumps
have been sold, and the Super MoneyMaker Plus is by far the
most commonly sold MoneyMaker pump (annex 18.2). Fig-
ure 18.4 details the number of pumps that have been sold
across the continent.
MoneyMaker pumps are developed through design cri-
teria consideration to meet the needs of the target market.
Final prototypes are field tested for long periods of time
(six months or more) with all feedback incorporated into
the final design. The first batch of pumps was field tested
through the private sector supply chain that stocked the
products and retailed them to the end users. Because
KickStart’s mission includes poverty reduction, an assess-
ment of the pumps’ impact is carried out at “zero age,” and
again in the 9th and 18th month after procurement. The zero
age survey provides the baseline data on which comparison
is made for the subsequent 9th-month and 18th-month sur-
veys. Manufacturing begins after the successful testing of
the pumps. Manufacturers are trained and provided with
parts produced by KickStart to ensure quality standards for
all the pumps. The pumps are supplied through distributors
and retailers for eventual sale to farmers.
KickStart embarks on intensive promotion and market-
ing campaigns to create awareness about and generate
demand for its products. Marketing and promotion of
high-quality products to poor people requires substantial
resources. For KickStart, this is possible through donor
and government support. Until a critical mass is reached
(15 to 20 percent of the market share), the tipping point is
not realized. However, once the tipping point is achieved,
operation costs are substantially reduced and the business
becomes sustainable and profitable and many more
distributors start selling the product.
MoneyMaker pumps are designed to last for three years
but some have, with good maintenance, operated for more
than six years. Critical in this good maintenance is the
replacement of piston cups every 3–12 months (depending
on hours of usage and the quality of water used).

Figure 18.4 Number of MoneyMaker Pumps in Various Countries, 1996–June 2009

50,000
Number of pumps sold

40,000

30,000

20,000

10,000

0
a

ia

da

rs
w

nd
ny

nd

bi
da

as
an

he
ala

an
m

ru
Ke

aF
Su

ga
nz

ot
Rw
Za
M

Bu
U
Ta

in

all
rk

l,
Bu

ta
/

To
ali
M

Source: KickStart sales data 1996–2009.

CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 325

 Figure 18.5 MoneyMaker Pump Supply Chain
Product
Production
R&D
Public sector
Continuous impact monitoring by KickStart
Source: KickStart International.
Table 18.3 Costs of Various Steps in the
MoneyMaker Pump Supply Chain
Supply Chain Step Cost ($)
Manufacturer 65
KickStart 72
Wholesaler 100
Retailer 115
Source: KickStart International.
MoneyMaker pumps supply chain
The MoneyMaker supply chain is a five-phase process
including research and development (R&D), production,
distribution, marketing and promotion, and sales and
backup services, as shown in figure 18.5. KickStart carries
out R&D and marketing and promotion directly. The
other phases are executed by the private sector with mon-
itoring by KickStart. Donor participation is very crucial to
carrying out product R&D and creating product awareness
because it enhances demand and consequently creates an
attractive business opportunity for the private sector.
Production, distribution and retailing costs, and profits
amount to $115 per pump for the Super Moneymaker
pump (table 18.3) and are borne by the farmers who pur-
chase the pumps. Other costs amounting to $264 a pump,
incurred during technology development, promotion, sales,
and impact monitoring, are currently borne by donors.
All the design work behind KickStart’s products is car-
ried out at Nairobi, while products are manufactured in
either Thika, Kenya; Arusha, Tanzania; or in China. Kick-
Start recruits and trains a network of distributors and
dealers (more than 150 dealers in Kenya alone) in cities,
towns, and small market centers. The pumps are sold, with
a markup of 7 percent to KickStart, to distributors in the
private sector through KickStart’s coordination, which
326 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA
Sales
and
Distribution Marketing &
backup
promotion
service
Private sector
maintains business accounts with each distributor or
dealer. Most of KickStart distributors are Agrovets shops
(sellers of agrochemicals and veterinary products) fre-
quently visited by farmers. Initially, KickStart provided
distributors with credit terms that required them to pay
KickStart once the consignment was sold. As demand for
KickStart products grew, the credit period was reduced to
30 days after receipt of consignment. KickStart also
requires that distributors be willing to demonstrate and
market the pumps.
KickStart identifies different marketing and promotion
strategies to create awareness among potential end users of
its products. These include above-the-line (those using
mass media) and below-the-line (nonmedia) promotions
to close sales. In Kenya, Tanzania, Mali, and Burkina Faso,
sales teams are trained to communicate the benefits of the
MoneyMaker pumps and promotion of farming as a busi-
ness to potential customers.
IMPACT OF MONEYMAKER PUMPS
ON DEVELOPMENT
Impact assessment methodology
Through its monitoring teams, KickStart assesses key indi-
cators of the impacts of the MoneyMaker range of pumps.
A range of data is collected to carry out this monitoring
process: the number of pumps manufactured and sold is
logged, purchasers’ details are recorded in a database, and
training recipients’ details are recorded. Later, KickStart’s
monitoring staff visit a random selection of distributers,
retailers, and purchasers to interview, administer question-
naires, and gather statistical data for technology impact
analysis. These socioeconomic surveys are conducted at
three points in time: at “zero age” (before use of the
pumps) and during the 9th and 18th month of usage. The In Kenya, socioeconomic surveys have shown a substan-
parameters monitored include pump ownership; pump tial increase in the area under irrigation at the household
management; and impact of the pumps on incomes, food level following the purchase of a MoneyMaker pump. At
security, livelihoods, and investment. the zero age survey an average of 0.02 hectares per house-
hold was under irrigation (typically bucket irrigation) but
Pump ownership and management rose to 0.2 hectares under hip pump irrigation, a nine-fold
average increase.
KickStart’s impact assessments (KickStart 2008) show that
88 percent of the people who purchase pumps are male.
Management of the pumps, however, is not carried out Impact on households
strictly by pump owners; rather, it changes with time. At
Although rural farmers are risk averse and have very little
zero age, 81 percent of the pumps are managed by the
cash to spare, demand for MoneyMaker pumps has
owners, a figure that fell to 46 percent at 18 months. On
increased steadily over the years, a trend that may arise from
the basis of gender, the surveys show that at zero age,
the change in lifestyles after purchasing the MoneyMaker
23 percent of pumps are managed by females, a figure that
pumps, as indicated by the profiles of four typical “farmer-
increased to 60 percent at 18 months. This is not surpris-
preneurs” shown in table 18.5.
ing since, as time passes, women’s interest in the pump
According to KickStart survey data, incomes in house-
grows because it provides a means of increasing income
holds using MoneyMaker pumps have risen from as little as
for families’ daily needs. Men, on the other hand, look at
$100 to more than $10,000 annually, in addition to extend-
other long-term high-capital investments that subse-
ing food availability periods from 3 months to 12 months,
quently generate higher returns. For example, farmer
thus eliminating food insecurity. The change in livelihood of
Samuel Ndungu (one of the four profiled below) opted to
the irrigating households is manifested in investment of the
leave his treadle pump to his wife while expanding his irri-
extra income in purchase of assets such as bicycles, motor-
gated enterprise with a motorized pump. This change in
ized pumps, dairy cows, modern houses, and land. Some
management shows how women have been empowered to
irrigating households have also expanded their enterprises
improve their livelihoods with MoneyMaker pumps.
to include dairy, poultry, small-scale maize milling, and
transport. With the increased incomes, payment of school
Impact on irrigated area fees for their children has eased, and the number of people
The area under irrigation using the pumps sold in Sub- employed on such farms has also increased. In addition, the
Saharan Africa between 1991 and 2009 is estimated to be MoneyMaker technology has encouraged the introduction
more than 31,000 hectares (table 18.4). of high-value crops, with their attendant improved agro-
nomic practices. And the technology has been frequently
adopted by neighboring farms, whose owners and workers
Table 18.4 Area under Irrigation Using the Pumps, have received on-farm training nearby.
and Crops Grown in those Areas in
Various Sub-Saharan Africa Countries,
Job Creation. Investment in agricultural water manage-
1991–2009
ment through small-scale irrigation creates jobs both for
Area being
the families who own the pumps and the labor they hire,
Country irrigateda Crops
because irrigation increases farm output and requires
Kenya 12,326 Horticultural crops
Tanzania 8,785 Horticultural crops more workers, especially where high-value crops are
Malawi 4,665 Maize, horticultural crops grown. Statistical analysis by KickStart (1991–2009), for
Sudan 1,531 Horticultural crops some Sub-Saharan countries, has shown that, on average,
Mali/Burkina Faso 1,433 Horticultural crops
Uganda 896 Horticultural crops when families started bucket irrigation, 0.17 jobs were cre-
Zambia 679 horticultural crops ated; after the acquisition of a MoneyMaker pump, the
Rwanda 540 Horticultural crops number of jobs created rose to 0.55 per household.1 Of
Burundi 284 Horticultural crops
Total 31,138
that 0.55, family members represented 0.51 and waged
labor represented 0.04, while distribution on gender basis
Source: Generated with KickStart sales data 1991–2009.
a. Area irrigated is generated using a conversion figure of 0.3 was 0.32 and 0.22 for men and women, respectively. The
hectares per pump at 80 percent pump utilization. jobs created included pump operation, land preparation,

CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 327

328

Table 18.5 Farmerpreneur Profiles before and after Investing in MoneyMaker Pumps

Name Samuel Mburu Ndungu Felix Muiruri Catherine Gwambie Mahamoud Guindo
Country Kenya Kenya Tanzania Mali
Age 45 34 40 48
Gender Male Male Female Male
Before After Before After Before After Before After
Farm enterprise Bucket irrigation Super MoneyMaker Bucket irrigation MoneyMaker Hip Rainfed and bucket Super MoneyMaker Bucket irrigation Super
and petrol pump Pump irrigation MoneyMaker
Income $100 $1,000–$2,000 $480 $10,440 — — $400 $700
($/year)
Food availability 3 12 3 12 3 12 3 12
(months)
Assets (house, 1 house, 1 cow, 2 houses, 2 cows, — — Chickens 1 modern house, — —
bicycle, farm 1 goat 7 goats, 3 bicycles, 2 businesses,
animals, etc.) one motorized chickens
pump
Land size (number 0.6 1.0 (0.4 being 0 0 0.8 1.2 0 0
of hectares) purchased as of
time of survey)
Land irrigated 0.4 MoneyMaker: 0.2 (rented) 0.8 (rented) 0.8 1.2 150 square meters 300 square
(hectares) 0.8 hectares meters
(0.4 rented);
motorized pump:
2 hectares
Family size 8 5 5 6
Number of Workforce has Workforce has 0 2 None None
employees grown fivefold doubled
on the farm
Crops French beans French beans, None French beans, Maize, beans Vegetables Vegetables Vegetables and
tomatoes tomatoes, baby fruits
corn, green
maize
Ease of Difficult Easy Difficult Easy Difficult Easy Difficult Easy
facilitating
children’s
education
Source: KickStart International.
— = Unavailable.
weeding, and harvesting. Pump operation represented the
largest proportion of the jobs created, with 33 percent of
the pumps being lent out to neighbors, generating 0.03
jobs at no cost. The Kickstart analysis also shows that
MoneyMaker pumps increased the number of jobs created
by more than 220 percent. In most countries, where many
youths are not currently engaged in productive activities,
irrigation provides an opportunity for a net increase in the
jobs created.
Income. In addition to job creation, MoneyMaker pumps
have a positive impact on farm incomes generated through
irrigation, as shown in table 18.2.
Poverty. Data (1991–2009) from Kenya, Tanzania, Mali,
Burkina Faso, and other countries show that 439,839 people
have been moved out of poverty (table 18.6). The poverty
reduction to pump ratio (number of people moved from
poverty per pump unit) varies but is clearly higher in Kenya,
Tanzania, Mali, and Burkina Faso than in other countries in
which the pumps are sold. This implies a higher efficiency of
pump utilization in the above four countries and is com-
mensurate with the level of technology promotion and
demonstration of on-farm profitability. These last two
factors are essential ingredients in high pump utilization
efficiency and need to be replicated in other Sub-Saharan
Africa countries to achieve similar success levels in those
locations.
IMPACT ON SUB-SAHARAN AFRICA
The impact of a technology depends on its adoption rate,
which is largely determined by the awareness created. Mar-
keting of the MoneyMaker technology requires a huge
financial investment that is not possible for many African
countries. Financial resources are particularly limited in Mali
and Burkina Faso, leading to low awareness levels and hence
low levels of impact compared with Kenya and Tanzania.
Despite the financial constraints, KickStart estimates that
between 1991 and 2009, more than 87,000 small-scale agri-
cultural enterprises were created using the MoneyMaker
range of pumps. These businesses generated new profits
totalling over $77.2 million annually and employing more
than 100,000 people. Each pump is used to irrigate 0.3
hectares of land (on average), and generates an average net
annual income of $1,100 a year for its owner, representing a
100–450 percent increase in the incomes that poor rural
people made from sale of crops before the adopting the
MoneyMaker technology. In Kenya by 2009 more than
12,000 hectares of land were estimated to be irrigated using
KickStart irrigation pumps. Irrigation investment costs less
than $60 per hectare annually compared with the conven-
tional sprinkler systems, which cost more than $600 per
hectare annually. In short, MoneyMaker pumps remove the
obstacle of huge initial capital investment for demand-
driven, small-scale irrigation development.
Kenya’s experience with MoneyMaker pumps
Thousands of entrepreneurial farmers in Kenya are now
irrigating with KickStart’s range of MoneyMaker pumps,
changing their small subsistence farms into vibrant com-
mercial enterprises. The average size of plots held by small-
holder farmers is between 0.2 and 1.0 hectare and generates
about 25–80 percent of total family income annually
(IPTRID 2005). When they use irrigation, farmers are typi-
cally able to grow and sell three to four high-value crops
annually. These “farmerpreneurs” in Kenya over 2007–09,
have increased their incomes ten-fold on average, and some
have made annual profits of up to $5,400.

Table 18.6 Number of People Using Kickstart Pumps Who Have Moved Out of Poverty in Select Sub-Saharan
Africa Countries, 1991–2009
Enterprises created People moved Poverty reduction
Country Pumps sold or transformed out of poverty to pump ratio
Kenya 51,357 40,720 203,600 3.96
Tanzania 36,603 29,495 147,475 4.03
Mali and Burkina Faso 5,970 4,673 23,365 3.91
Other countriesa 43,676 13,079 65,399 1.50
Total 137,606 87,967 439,839 3.20

Source: KickStart 2008.

Note: Here, poverty is defined as an income of less than $1 a day.
a. Burundi, Democratic Republic of Congo, Malawi, Mozambique, Rwanda, Sudan, Uganda, and Zambia.

CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 329

 Distribution and Utilization. The distribution of the
MoneyMaker range of pumps in various regions of Kenya is
shown in figure 18.6. This distribution reflects the agricul-
tural potential in the region, the ability of farmerpreneurs to
purchase the pumps, and the availability of surface and
shallow groundwater sources.
Wealth Creation. According to KickStart data, the aver-
age net annual income per family using the MoneyMaker
pumps in Kenya ranges from $800 to $10,440. As of
December 2005 about 33,000 pumps had been purchased
by farmers and were generating household income
totalling $31.4 million (K Sh 2.2 billion) annually. Impact
monitoring survey data showed that 91 percent of the pur-
chased pumps generated total income of $93.8 million (K
Sh 6.7 billion) during their life span of three years. The
other 9 percent are pumps that were not used for various
reasons such as water source being too deep for the pump
to reach. Table 18.7 shows the income generated by each
type of MoneyMaker pump.
Income and Employment Creation. In most parts of
rural Kenya, as elsewhere in Sub-Saharan Africa, time is
an available resource that is very often underutilized
because opportunities for productive employment activi-
ties are lacking. KickStart’s ranges of MoneyMaker pumps
are designed to create opportunities for rural people to
more fully use their available time on productive activi-
ties. In Kenya, the average annual household income rose
from $3,465 (K Sh 242,529) before the pump to $4,700
(K Sh 328,973) after acquisition of the pump (KickStart
sales data 1991–2009). Although total household income
rose from several sources, including farming, employ-
ment, pension, remittances, businesses, and collection of
rent, income from irrigation contributed the highest pro-
portion of the total, at 47 percent. KickStart’s monitoring

Figure 18.6 Regional Distribution of MoneyMaker Pumps Sold in Kenya, 2007–09

2,500

2,000
Pumps sold

1,500

1,000

500

0
al

st

rn
ob

nz

lle
er

er
tr

oa

te
air

ya

Va
en

st

st
C

es
Ea

Ea

N
N
C

W
ft
Ri
th
or
N

2007 2008 2009

Source: KickStart.

Table 18.7 Wealth Creation with MoneyMaker Pumps in Kenya

Pumps purchased Annual income Cumulative income over
Irrigation technology 1996 through 2005 ($ millions) three years ($ millions)
MoneyMaker Suction Pump 3,305 2.31 6.93
Super MoneyMaker 13,657 17.44 52.31
Super MoneyMaker plus 7,401 9.45 28.35
MoneyMaker Plus 6,639 2.00 6.00
Hand Pump 2,221 0.23 0.23
Total 33,223 31.27 93.82

Source: IPTRID 2005.

330 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA

surveys (2008) also show that Super MoneyMaker pump
irrigation over an 18-month period on average increased
pumping and watering jobs by 465 percent in Kenyan
farms. This is in addition to increased jobs in other crop
production activities. Based on the number of pumps
sold, the area irrigated and hence the number of jobs
created, Kenya has made the largest gains of all the coun-
tries in which MoneyMaker pumps have been sold.
Assets, Housing, and Investment. KickStart surveys
show that in Kenyan households using MoneyMaker
pumps, the number of local cattle owned declined by
31 percent, whereas the number of the more expensive
dairy cows increased by 7 percent, indicating availability of
surplus resources available to purchase them. Dairy cows
produce more milk than the local cattle, thus increasing
family income. There was also more fodder available to
feed the dairy cows, which demand more fodder than local
cows. Similarly, there was an increase in the number of
donkeys and carts in households using MoneyMaker
pumps, indicating the demand for extra transport with the
use of MoneyMaker technology. In addition, households
increased their investments in businesses (shops, transport,
poultry, and dairy) by 32 percent. KickStart surveys also
show improvement in housing among households in Kenya
Box 18.1 Money Maker Success Stories
Daniel Karanja Njenga and Nancy Gathoni
in Kenya
While living in an internally displaced person camp
after their home and farm were looted during Kenya’s
postelection violence in March 2008, Daniel Karanja
Njenga and his wife Nancy Gathoni participated in a
promotional demonstration of the MoneyMaker Hip
Pump. “When I saw the . . . demonstration and heard
about it on the radio,” Daniel said, “I knew it was the
answer to how we could earn an income quickly and get
back to farming.”
Daniel’s first purchase with the relief funds provided
to him by the Kenyan government ($130) was the Mon-
eyMaker Hip Pump and hoses. The manually operated
pump is lightweight, costs less than other pumps, and is
easy to use. It does not require electricity or fuel. With it,
Daniel was able to irrigate his small 1/8th-acre plot and
grow crops during the dry season when most farms are
bare. Daniel and Nancy now earn a decent living selling
their sukuma wiki (kale), a staple food for Kenyans. They
CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA
after they started MoneyMaker pumps. In 2008, the
incidence of “improved” houses—those constructed from
stone and timber—increased by 38 percent and 10 percent,
respectively, demonstrating investment in better housing
and therefore improved livelihoods.
Food Security. Food security exists when people have
physical and economic access to food that is sufficient
and nutritious enough to meet their dietary needs and
preferences for an active and healthy life. Food security
can be achieved through producing one’s own food (self-
sufficiency), having sufficient income to purchase food, or
a combination of both. KickStart surveys (2008) show
that households using the Super MoneyMaker pump
increased their food surplus by 35 percent in the
18 months following the adoption of the technology.
DRIVERS OF SUCCESS
MoneyMaker success stories are numerous. Three such sto-
ries from Kenya, Tanzania, and Burkina Faso are presented
in box 18.1.
In developing its irrigation pumps, it is KickStart’s pol-
icy to ensure four main outcomes: achieving the highest
impact in the shortest time possible; achieving a
are also helping hungry neighbors and friends struggling
to get back on their feet after the violence. Daniel has
plans to expand his plot to grow cabbages and tomatoes
and purchase a dairy cow.
Daniel and Nancy are two of many other farmers in
Kenya who are becoming successful businessmen and
women able to feed their families, pay school fees, and
medical expenses. According to KickStart’s research, an
average farmer can make about $120 a month selling
crops produced using the MoneyMaker pump.
In Kenya MoneyMaker pumps are also providing
hope to people facing a tough combination of challenges
in recent years: postelection violence, escalating food
prices, a difficult economic situation, and continued
high levels of poverty.
Catherine Gwambie in Tanzania
Catherine Gwambie and her husband, Hawzi Mwami,
are an entrepreneurial couple from Tanzania who had
dreams of becoming successful shop owners in Dar es
331
 Box 18.1 (continued)
Salaam. They farmed in their native Kigoma, growing
and selling maize and beans to save enough to open a
shop selling household supplies.
Although the shop was reasonably successful, it did
not generate as much income as the couple needed to
support their family. Hawzi decided to buy land on
which he could raise chickens and Catherine could
start growing vegetables for sale. It was a good busi-
ness, but bucket irrigation on the land took a lot of
effort.
In early 2007 Catherine heard an advertisement for
the Super MoneyMaker pump on the radio. She excit-
edly told her husband about this new pump that was
affordable and made irrigation easier and quicker.
Hawzi was not convinced. Catherine, however, insisted
that the pump would make her life easier and decided
to use her own money to purchase it. Together, the
Mwamis went to the Kariakoo market in Dar es Salaam
to buy a Super MoneyMaker.
The pump increased Catherine’s productivity so
much that she expanded to another plot and now
employs her daughter and young sister. With the extra
income they are now earning, the Mwamis have plans to
send three of their young children to good secondary
schools and to build a nicer house for their family.
Source: Authors.
cost-effective program; ensuring that the income-generat-
ing process is self-sustainable, and ensuring that the
process if scalable and replicable in other areas.
KickStart employs six approaches to achieve its devel-
opment goals, as described below. In addition, KickStart’s
success has been driven by the desire of farmer to produce
on-farm profits and get out of poverty through adoption
of appropriate and affordable technology
Market Research. KickStart’s technological gap studies
identify profitable small enterprises that can be estab-
lished by local entrepreneurs with limited capital invest-
ment and determine the technological and socioeconomic
requirements of those opportunities. The studies evaluate
raw materials, competing products, potential market
demands, and constraints and opportunities for small
enterprises.
332 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA
Hawzi now freely admits that his wife was right about
the pump, and between their two businesses, they see a
bright future for their family.
Mahmoud Guindo in Mali
Mahmoud Guindo, a 48-year-old farmer with a wife and
four children, had long struggled to make ends meet on
credit. He moved from his home in Dogon County to
Bamako, where he was employed as a security guard
earning $400 annually. In an effort to boost his income,
he began farming a 150-square-meter plot, which was
still inadequate to meet his family’s needs.
To increase his annual income, Mahmoud wanted to
build a bigger garden, but he was sceptical about how he
would water a larger plot of land. After seeing an adver-
tisement for KickStart’s MoneyMaker pump on television
in 2008, Mahmoud wanted to buy it. He did not have
enough money, however, so he approached his boss for a
loan. Both men agreed it was a tangible asset that would
provide a profitable and quick return on investment.
Since buying the pump in October 2008, Mahmoud
has almost doubled his annual income, from $400 to
$700, by selling fruits and vegetables. The additional cash
flow is allowing him to pay off some debts while simul-
taneously providing sufficient food for his family.
Designing New Technologies and Business Packages.
KickStart designs and develops the tools, equipment, man-
uals, and business plans required for establishing small
enterprises. It also designs the production protocols and
quality control procedures required for manufacturing
new pumps.
Training Manufacturers to Produce MoneyMaker
Pumps. KickStart trains private manufacturers to set up
assembly lines and quality control mechanisms to mass-
produce its machines and tools.
Promotion of MoneyMaker Pumps. KickStart promotes
the MoneyMaker and works with the private sector to ensure
that pumps are well known and easily available to small-scale
investors, reducing product risk for entrepreneurs. The risk
to entrepreneurs is further reduced by the fact that the
pumps are guaranteed for a period of one year following
purchase. Farmers have the right return their pumps if their
wells are too deep for the operation of the pump or if the
pumps break down due any manufacturing fault.
On-farm Profitability. The enhanced water utilization
that has come with the MoneyMaker range of pumps has
increased farm incomes, a major incentive for practicing
and potential farmers.
Farmers’ Need to Get Out of Poverty. Despite being
hardworking and dedicated, farmers in Sub-Saharan Africa
have long been seeking a means to escape poverty. The
MoneyMaker technology has accorded them an affordable
opportunity to do just that.
Partnership with Donors. KickStart raises funds for
research and development, marketing, and awareness pro-
motion of MoneyMaker technologies and associated busi-
ness plans. KickStart’s donors include a large number of
public and private donors such as the Bill and Melinda
Gates Foundation, the Rockefeller Foundation, and the
David and Lucille Packard Foundation.
SCALABILITY AND TRANSFERABILITY
OF THE SUCCESS
The scalability and transferability of KickStart depends cru-
cially on the existence of three things: a sustainable supply
chain involving manufacturers, distributors, and retailers;
use of appropriate marketing technologies until the tipping
point is reached; thereafter, the technology is self-propagating;
and policy changes to encourage entrepreneurship and
removal of handouts of technologies.
Maximization of the use of the technology is essential. For
example, the Super MoneyMaker is currently used to irrigate
an average 0.3 hectares, despite being designed to irrigate
0.8 hectares. This limitation is attributed to the limited land
available to the farmer. Participatory usage, in which farmers
share a single pump, can also increase the use of the pumps
and the amount of land area being irrigated. Additionally,
expansion of the pump’s utility from sprinkler irrigation to
include drip irrigation would enhance the water and energy
use efficiencies and therefore increase the land coverage. In
Naro Moru, Kenya, for example, MoneyMaker pumps are
used to pump water from shallow wells into raised storage
tanks that feed into drip irrigation systems. This arrange-
ment has the effect of increasing the area under irrigation.
CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA
If each family in Kenya with access to water from
underground storage, rivers, streams, ponds, or dams used
a treadle pump and irrigated 0.3 hectares of land, for
example, that country would not be experiencing the food
shortages and hunger it has recently. Assuming water
accessibility, an irrigation potential of 540,000 hectares,
and a 20 percent share of this irrigation potential for the
MoneyMaker technology, Kenya would need about
360,000 of the MoneyMaker range of pumps. Based on
KickStart’s sales figures, as of 2005, only 9.2 percent of the
potential number of pumps had been sold, and 14.2 per-
cent by October 2009 (KickStart sales data 1991–2009).
Additionally, KickStart aims to increase access to 20 percent
of the potential irrigation market in Kenya, after which the
technology will be self-propagating. After the tipping point
is reached and assuming 80 percent pump utilization effi-
ciency, 17,280 hectares would be under irrigation through
the MoneyMaker range of pumps.
In Sub-Saharan Africa, more than 87,967 small agricul-
tural enterprises had been created by 2009 using Money-
Maker pumps. These businesses generate $81 million each
year and offer 93,462 new jobs annually. This new wealth
has helped move 439,839 people out of poverty. A further
indication of the benefits of irrigation is the establishment
of small-scale businesses within the vicinity of nearby vil-
lage markets, thus creating indirect employment. This has
led to improved access to nutrition, education, health, hous-
ing, and welfare services for farming communities.
KickStart’s systematic, replicable method of measuring
the impacts of its products is an aspect of the organization’s
success that could be replicated by other organizations.
Every product comes with a one-year guarantee, and every
buyer fills out a guarantee form when the product is pur-
chased. The guarantee reduces the perceived risk of buying
the product, and the forms give KickStart a database of all
pump owners.
Scaling up the MoneyMaker technology success in Sub-
Saharan Africa should involve technology promotion,
demonstration of on-farm profitability and a sustainable
supply chain, and high levels of efficient pump use. The
still-low pump utilization needs to be addressed through
on-farm pump utilization research that looks at, for exam-
ple, shifting from direct irrigation pumping to indirect
through storage tanks, or changing the application
method from sprinkler to drip. All these require enhanced
budget levels similar to or greater than the current promo-
tional budgets in Kenya and other African countries. In
addition to the aforementioned interventions, government
333
 policies should shift to favor accessible small-scale irriga-
tion technologies.
LESSONS LEARNED FROM
MONEYMAKER PUMPS
Prospects
In Sub-Saharan Africa, intensification, rather than expan-
sion, of cropped areas, is key to agricultural growth and
poverty reduction. One way of achieving this is introduc-
tion of simple innovative technologies such as Money-
Maker pumps for agricultural water management. Through
irrigation, farmers can diversify into high-value horticul-
tural crops and fodder crops.
The concept of farmer entrepreneurs, in which agricul-
tural enterprises are run as viable businesses, is now fully inte-
grated as policy in Kenya and needs to be introduced in other
African countries. As KickStart’s products have shown, when
poor people have access to technology to generate wealth,
they use that technology quite effectively to move themselves
out of poverty. Technology acceptance by farmers and a sus-
tainable supply chain is assured through the KickStart model.
A study (IPTRID 2006) on the experience with treadle
pumps, introduced in West Africa by an international NGO,
Enterprise Works, found the need for improved after-sales
service and a coordinating agency with a longer-term pro-
gram, such as the sort KickStart provides in East Africa. The
study recommended the adoption of the KickStart model to
ensure the sustainability of the West African program.
334 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA
Technology infusion
The participatory approaches to technology infusion
employed in the KickStart model involve marketing
through demonstration and competitions. Marketing of
technology is important in creating and enhancing adop-
tion of technology. Creation of organized marketing
groups and structures is important in poverty reduction.
Cultural and gender sensitivity in technology develop-
ment and infusion is an important consideration in tech-
nology. Technology evolution changes are driven by users
of the technology. For example, in response to user
demands, the Super MoneyMaker pressure pump, which
facilitated sprinkler irrigation, was developed to replace
the original suction-only pump and was quickly adopted
by farmers.
Challenges
Despite its efforts and success, KickStart has not been able
to design the ideal pump that combines high performance
without maintenance irrespective of the intensity of use,
terrain in which it is operated, and water quality.
The low pump utilization efficiency, irrigating on
0.3 hectares of the possible 0.8 hectares, could be attributed
to direct irrigation pumping, which causes fluctuating
sprays. It may be possible, however, to improve pump uti-
lization efficiency through indirect irrigation pumping
through storage tanks or using the pumps in combination
with more efficient water application methods.
Annex 18.1 Economic Analysis of Pumps Given Away versus Sold
Giveaway KickStart
Item model selling model Remarks
Donor funds $2 million $2 million Start with the same amount of funds.
Costs per pump $290 $257 Each pump is $33 cheaper to sell than to give away.
Manufacturing 65 65 Why? It costs $65 to manufacture each pump; they are then sold to a wholesaler at
Revenue 0 $–72 $72, giving revenue of $7 per unit. This earned income helps support the
Distribution $85 0 organization.
Promotion and sales 0 $124 Costs of administration, fundraising, technology development, and impact monitoring
Admin, technology $140 $140 are the same in either model. The giveaway would not require marketing and
development, promotion but would incur costs to distribute the pumps. Staff would be needed
impact monitoring, to coordinate, and vehicles and fuel would be needed to transport the pumps.
fundraising KickStart would also need to provide hoses or the pumps would be useless. In
the KickStart model, distribution costs are handled by the wholesaler.
Total units 6,897 7,782 Donor funds divided by cost per pump. The KickStart model puts nearly 900 more
distributed pumps in the field.
Percent used in 30% 80% KickStart’s goal is not to distribute pumps. Rather, it is to help people create small
enterprise enterprises. It is here where the difference in the two approaches becomes
Number used in 2,069 6,226 apparent. When a person makes an investment, he or she is committed to making
enterprise a better future. Impact monitoring data indicates that 80 percent or more of
pumps bought are used to create jobs and income. The same research shows less
than 30 percent of pumps given away are used to create a business. The KickStart
model creates three times more small businesses than the giveaway model.
Average new profits $1,100 $1,100 These businesses will generate an average of $1,100 in new profits and wages each
and wages year.
Net present worth $9,103,600 $27,394,400 Here is an estimate of the new profits and wages these new businesses will create
over 4 years in four years. The KickStart model leverages $2 million in donor funds into more
than $27 million in profits and wages for these families—three times more than
the giveaway model. This is money that is spent locally, supporting other business
and stimulating the local economy.
Number of people 10,345 31,130 KickStart estimates that each enterprise supports a family of five. The KickStart
moved out of poverty Model moves more than 20,000 more people out of poverty, and it does it at
Cost to move one $193 $64 one-third the cost per person of the giveaway model.
person out of poverty
Source: KickStart International.

Annex 18.2 Number and Type of Pumps Sold in Various Countries, July 1991–November 2009
Mali/
Burkina Other
Pump type Kenya Tanzania Malawi Sudan Faso Uganda Zambia countries
MoneyMaker 3,305 438 0 0 0 212 0 95
Super MoneyMaker 13,683 3,837 0 80 0 539 1 857
Super MoneyMaker Plus 16,985 26,988 18,503 6,138 5,609 2,885 2,368 7,709
MoneyMaker Plus 7,674 264 0 141 0 0 0 30
MoneyMaker Hand Pump 2,217 431 30 0 0 0 0 50
MoneyMaker Hip Pump 7,493 4,645 905 120 361 97 462 2,454
Total 51,357 36,603 19,438 6,379 5,970 3,733 2,831 11,295
Grand total 137,606
Source: KickStart International.

CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 335

 NOTE
1. The analysis defines a job as an economic activity that
occupies an individual for five hours daily for 150 days
annually.
BIBLIOGRAPHY
Aquastat. http://www.fao.org/nr/water/aquastat/main/index
.stm.
Arbache, J., and J. Page. 2008. “Hunting for Leopards: Long-
Run Country Income Dynamics in Africa.” Policy
Research Paper 4715, World Bank, Washington, DC.
Brookings Institution. 2007. “Panel 5: Africa’s Economic Suc-
cesses: What’s Worked and What’s Next.” Brookings Blum
Roundtable 2007. Brookings Institution, Washington, DC.
FAO (Food and Agriculture Organization). 2005. “Irriga-
tion in Africa in Figures: Aquastat Survey 2005.”
http://www.fao.org/nr/water/aquastat/regions/africa/
index.stm.
IFAD (International Fund for Agricultural Development).
2005. “Agriculture Development in Republic of
Mozambique.” Agriculture Support Programme For-
mulation Report, Working Paper 2, International Fund
for Agricultural Development, Rome.
IPTRID (International Programme for Technology and
Research in Irrigation and Drainage). 2005. “Kenya
Impact Case Study Report.” IPTRID, Rome.
IPTRID. 2006. “Treadle Pump Dissemination and Adoption
in West Africa: Performance Problems and Prospects.”
GRID Network 24. Rome.
KickStart. 2008. “The Super MoneyMaker Pump: The
18-Months Impact Assessment Report.”
Kidane W., M. Maetz and P. Dardel. 2006. “Food Security
and Agricultural Development in Sub-Saharan Africa:
Building a Case for More Public Support.” Main Report.
FAO Regional Office for Africa, Harare.
336 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA
Morris, M., H. Binswanger-Mkhize, and D. Byerlee. 2009.
Awakening Africa’s Sleeping Giant: Prospects for Commer-
cial Agriculture in the Guinea Savannah Zone and
Beyond. Washington, DC: World Bank.
Msangi, S., and M. Rosegrant. 2005. “World Agriculture in a
Dynamically Changing Environment: IFPRI’s Long-Term
Outlook for Food and Agriculture under Additional
Demand and Constraints.” Paper written for the FAO and
United Nations Economic and Social Development
Department Expert Meeting on “How to Feed the World
in 2050.” International Food Policy Research Institute,
Washington, DC.
NEPAD (New Partnership for Africa’s Development) and
FAO (Food and Agriculture Organization). 2004.
“Government of the Republic of Mozambique: National
Medium Term Investment Programme.” ftp://ftp.fao.org/
docrep/fao/007/ae415e/ae415e00.pdf.
Rosegrant, M., and N. Perez. 1997. “Water Resources
Development in Africa: A Review and Synthesis of
Issues, Potentials, and Strategies for the Future.” EPTD
Discussion Paper 28, International Food Policy
Research Institute, Washington, DC.
Tyler, G. 2009. “All-Africa Review of Experiences with
Commercial Agriculture: The African Sugar Industry—A
Frustrated Success Story.” Background Paper for the
Competitive Commercial Agriculture in Sub-Saharan
Africa (CCAA) Study. World Bank and FAO, Washington,
DC. http://siteresources.worldbank.org/INTAFRICA/
Resources/257994-1215457178567/Ch6_Sugar.pdf.
UNIDO (United Nations Industrial Development Organi-
zation). 2009. Industrial Development Report 2009:
Breaking In and Moving Up: New Industrial Challenges for
the Bottom Billion and the Middle-Income Countries.
Vienna: UNIDO.
World Bank. 2000. Can Africa Claim the 21st Century?
Washington, DC: World Bank.
PA R T V

Engaging the Private Sector to Upgrade

Infrastructure
 CHAPTER 19

ICT in Sub-Saharan Africa: Success Stories

Kaoru Kimura, Duncan Wambogo Omole, and Mark Williams

s recently as 1998 few people in Sub-Saharan

A
Voice networks
Africa had access to a telephone, and even fewer
Voice networks are expanding rapidly, and the number of
had access to computers. Since then, this situation
people using the networks is growing. Between 1998 and
has changed out of all recognition. Africa has experienced a
2008 the number of mobile cellular subscribers in Sub-
continent-wide revolution in the growth of the information
Saharan Africa leaped from about 4 million to 259 million
and communication technology (ICT) sector. Networks
(figure 19.1)
have expanded, services have been made available on a mass
Innovations in the way services are delivered have also
market basis, and prices have fallen. ICT is now an everyday
made ICT more accessible to customers in Africa. Prepay-
service for many Africans, affordable to the majority rather
ment has allowed customers to control their expenditures
than the privileged few. This growth continues today as
and avoid regular monthly subscription payments. African
companies innovate, bringing new products and services to
operators have also led the way in selling prepay credits in
the market.
very small units, making calling accessible to large numbers
This chapter examines the ICT sector in Africa. The first
of people, including the poor.
section describes the expansion of voice networks and
A combination of new technologies and sound policy
broadband Internet. The second section identifies the key
choices, including market liberalization, by governments
factors that have contributed to the expansion of the sector.
has triggered private investment, which has laid the founda-
The third section analyzes the impact of ICT on a variety of
tions of the connectivity revolution in Africa. The global
areas, including mobile banking, telemedicine, and agricul-
standardization of mobile telephone network technology
ture. The last section examines the role of the World Bank
has created international competition in both network
Group, to date and in the future.
equipment and customer devices, pushing down costs and
lowering prices. The liberalization of markets and the estab-
lishment of sound regulatory frameworks have allowed pri-
EXPANSION OF THE SECTOR
vate investors into the market, driving the expansion of
The telecommunications sector in Africa has expanded telecommunications networks. Between 1998 and 2008, total
rapidly since 1998, with both an increase in network and private sector investment in telecoms in Sub-Saharan Africa
services provided. At the same time, prices have fallen every- reached $49 billion (World Bank 2010a), mainly in the
where, bringing telecommunications within the economic mobile market (figure 19.2). This investment has resulted in
reach of the majority of Africans. a rapid increase in the proportion of the population covered

339
 Figure 19.1 Mobile Cellular Subscription in Sub-Saharan Africa, 1998–2008

300 70

258.5
250 60

Subscribers per 100 people

50
200 184.3

40
Million

150 135.3
30

100 90.7
20
54.2
50 36.3
25.1 10
17.0
6.5 1.011.4 1.7
0.6
0 0
98

99

00

01

02

03

04

05

06

07

08
19

19

20

20

20

20

20

20

20

20

20
Mobile cellular subscriptions in Sub-Saharan Africa (left axis)
Mobile cellular subscriptions in Sub-Saharan Africa per 100 people (right axis)
Mobile cellular subscriptions in world per 100 people (right axis)

Source: World Bank 2010b.

Figure 19.2 Telecommunications Investment in The success of mobile communication has been very
Sub-Saharan Africa, 1998–2008 scalable, with large and small countries experiencing rapid
increases in network coverage and access. Both Nigeria, a
12.0 country of about 150 million people, and Rwanda, a coun-
10.0
try of about 10 million people, have experienced steady
growth in network coverage (figure 19.4).
Billions of dollars

8.0 At the same time as communications networks have

6.0
been expanding across Africa and the number of sub-
scribers has risen, the price of telecommunications services
4.0 has been coming down. The cost of mobile communica-
2.0
tions fell by nearly 50 percent between 2000 and 2009. By
2010 the average price of a mobile call in Africa was about
0.0 $0.10 a minute (AICD forthcoming), and prices are con-
98

99

00

01

02

03

05

06

07

08

tinuing to fall as competition intensifies.

0
19

19

20

20

20

20

20

20

20

20

20

Source: World Bank 2010a.

Broadband internet
by mobile networks, which increased from about 10 percent Globally, broadband Internet is becoming an important
in 1999 to 60 percent in 2009. More than 90 percent of Sub- part of national economies. Africa lags far behind most
Saharan Africa’s urban population now lives within range other parts of the world in access to broadband Internet.
of a mobile signal, a remarkable transformation in only Broadband penetration in the region was 6.5 percent in
10 years (figure 19.3). 2008, one-quarter of the world average (ITU 2010). In 2008

340 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES

Figure 19.3 Mobile Coverage in Rural and Urban Sub-Saharan Africa, 1999–2009

100
90
80
70
60
Percent 50
40
30
20
10
0
Ju 9
Jan 9
Ju 0
Jan 0
Ju 1
Jan 1
Ju 2
Jan 2
Ju 3
Jan 3
Ju 4
Jan 4
Ju 5
Jan 5
Ju 6
Jan 6
Ju 7
Jan 7
Ju 8
Jan 8
9
-9
l -9
-0
l -0
-0
l -0
-0
l -0
-0
l -0
-0
l -0
-0
l -0
-0
l -0
-0
l -0
-0
l -0
-0
Jan

Total Urban Rural

Source: Authors based on data from GSM Association.

Note: Figures show connection to GSM network.

Figure 19.4 Mobile Network Coverage in Nigeria and Rwanda, 2002–08

100
90
80
Percent of population

70
60
50
40
30
20
10
0
02

03

04

05

06

07

08
20

20

20

20

20

20

20

Rwanda Nigeria

Source: ITU 2010.

fixed broadband Internet cost $100 a month in Sub- KEY SUCCESS FACTORS
Saharan Africa, more than three times the world average
Many factors have contributed to the success of the ICT sec-
of $31.40 (ITU 2010) (figure 19.5).
tor in Africa. The most important of these has been the con-
This situation is starting to change, as broadband Inter-
tinentwide change in sector policy from one based on
net becomes more widely available and prices begin to fall.
monopoly provision of ICT services to a privately owned,
Nigeria, Kenya, and South Africa are leading the way in
competitive market. The vast majority of countries in Africa
Africa, with rapidly rising numbers of broadband sub-
have liberalized their telecommunications markets, and that
scribers, mainly through wireless handsets. In most African
liberalization has spurred private investment into infrastruc-
countries, however, broadband remains beyond the reach of
ture and created competition between operators, leading to
the majority of the population.
expanded availability of services and reduced prices.

CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 341

 Figure 19.5 Monthly Price of Broadband Internet, by World Region, 2008

120 120

Broadband price as percent of monthly per capita income

100
100 100
101

Monthly broadband internet price (dollars)

80 80

60 60

40 34 40
31

22 23 23 21
20 20
26

9 3
5 8 4
0 0
cif d

As d

be d

Af and

ia

ld
ric
Pa an

al an

ib an

As

or
ic

ia

an

Af
ric

W
ia

tr e

ar a

th st

h
C ic
en p
As

or a

ut

n
C ro

e er

N eE

ra
So
Eu

th Am
st

ha
dl
Ea

id

Sa
tin

b-
La

Su

Fixed broadband internet price ($ per month), left hand side

Broadband price percent of per capita income (monthly), right hand side

Source: ITU 2010.

Many governments have also privatized the incumbent The success of the ICT sector in Africa is also attribut-
telecommunications operator. Privatization is significant able to innovation within the industry itself. Technologies
because it removes one of the main distortionary factors and business models have been adapted to the African
in the market and allows the government to carry out market, improving efficiency and tailoring services to the
its main role of regulating the market without any conflict specific needs of customers in the region. Finally, the
of interest. political economy of the telecommunications sector has
These policies have resulted in the rapid growth of the affected government’s willingness to undertake sector
mobile telecommunications segment of the market. Broad- policy reforms.
band Internet has been much slower to take off in part
because of the inadequacy of certain segments of the
Creating competitive markets
broadband network infrastructure, particularly access to
the international communications networks. These bottle- Competition in the telecommunications market was intro-
necks have been constraining growth of broadband Inter- duced in Africa in the late 1990s, when mobile networks
net, limiting its availability, and making it prohibitively were first established there. All but a handful of countries
expensive. Recently, the growth of the submarine fiber- in Sub-Saharan Africa have introduced competition in their
optic cable infrastructure has dramatically increased the mobile markets, and more than half have more than two
availability of international bandwidth on the continent, mobile operators (ACID forthcoming). This increase in
and competition between cable providers has led to rapid competition has been one of the primary drivers of the
declines in prices. sector growth and network expansion.

342 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES

 Competition takes time to develop. Countries that
reformed early have experienced better sector performance
than late reformers. Performance also improves as competi-
tion intensifies. Mobile subscription penetration rates in
Africa grew by less than 1 percentage point a year between
the establishment of the first mobile network and the entry
of the second operator. By the time a fourth operator enters
the market, the subscriber penetration rate grows by 2–4
percentage points a year (AICD forthcoming).
The way in which liberalization takes place also has an
impact. In Nigeria, for example, the government awarded
the first three GSM (Global System for Mobile communica-
tions) licenses simultaneously, spurring a race to expand
coverage and gain market share. Nigeria now has five mobile
operators and is considered one of the most competitive
markets in the region. With a mobile subscriber penetration
rate approaching 50 percent, it is also one of the best con-
nected countries in Africa.
Competition has been the primary driving force behind
mobile network expansion and, more recently, declines in
the price of calls. A recent study indicates that competition
could potentially continue to drive this expansion until net-
works cover 90 percent of Africa’s population (AICD 2010).
Privatizing state-owned enterprises
State ownership of telecommunications operators distorts
markets and adversely affects competition. Governments
often give state-owned operators preferential treatment
by carving out areas of exclusivity for them in the market
or failing to enforce regulatory rules such as interconnec-
tion payments. At the same time, these operators often
find it difficult to obtain the necessary capital and skills to
Box 19.1 Privatization and Sector Reform in Mali
Mali’s state-owned telecommunications operator,
Société des Télécommunications du Mali (SOTELMA),
was established in 1989 as the state-owned monopoly
operator. Despite the continuous growth in the mobile
market in Mali—led by the introduction of competi-
tion, in the form of a second GSM operator, Ikatel in
2003—the fixed-line market remained underserved and
dominated by SOTELMA.
In 2009 SOTELMA was successfully privatized, with
a 51 percent stake going to Maroc Telecom. At the same
time, a new legal framework and universal access strat-
egy were adopted. The telecommunications penetration
compete effectively with the private sector. The best long-
term solution to these problems is to sell state-owned
enterprises to create a level playing field for all players in
the market.
Realizing this, governments across the region have priva-
tized their state-owned telecommunications operators. By
2010, 28 African countries had gone through this process
(see annex table 19A.1). Mali completed the privatization
of its state-owned telecommunications operator in 2009
(box 19.1).
Overcoming infrastructure bottlenecks
Most mobile operators have built their own end-to-end net-
works, overcoming infrastructure bottlenecks along the way.
As the focus of the ICT industry in Africa shifts toward
broadband Internet, network infrastructure requirements
are changing. Broadband requires networks capable of han-
dling much higher volumes of traffic than that generated by
voice services. As much of the traffic generated by broad-
band crosses international borders, it also requires more
international network infrastructure. This type of network
infrastructure has traditionally been underdeveloped in
Africa. The lack of suitable infrastructure has become a sig-
nificant bottleneck, contributing to the slow growth of the
Internet in the region.
This situation began to change in 2009 with the develop-
ment of undersea fiber-optic cables connecting Africa to the
global communications networks. By 2010 Sub-Saharan
Africa had 12 operational cables, and another 5 were under
construction. The operational cables have a combined
capacity of more than 12 terabytes per second (tbps). A total
of $1.7 billion is being invested in the 5 undersea cables
rate rose to 40 percent, and Maroc Telecom is now
focusing on revitalizing the fixed-line network and
rolling out fixed broadband Internet.
This process was supported throughout by the
World Bank, through a technical assistance project
covering the privatization, the reform of the legal and
regulatory framework, and capacity building for the
government institutions. A World Bank team contin-
ues to work with the government of Mali to further
liberalize the telecommunication market and improve
connectivity by developing regional links to neighbor-
ing countries.
CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 343
 currently under construction, which will bring an addi- development of the Internet in Africa. Development of
tional 9 tbps of capacity to the region (figure 19.6). these networks has been much slower, although the recent
The impact of this growth is already being felt. In East upsurge of interest by African telecommunications compa-
and Southern Africa, the SEACOM, TEAMs, and EASSy nies in broadband has prompted new investment in
undersea cable projects were all operational by 2010. domestic fiber-optic networks. For the foreseeable future,
Wholesale prices for international bandwidth fell, allowing this interest is likely to focus on servicing the profitable
operators to provide better-quality service, usually in the intercity routes rather than connecting smaller towns and
form of faster download speeds or larger download caps. rural areas. A broadband infrastructure bottleneck is there-
The arrival in 2010 of new undersea cables along the west fore likely to remain in these areas for some time to come
coast of Africa (Glo-1 and MainOne) together with the (Williams 2010).
expected arrival of two more cables (ACE and WACS) in
2011 and 2012 is likely to have a similar effect on markets on
Fostering innovation
the other side of the continent. The international broad-
band infrastructure bottleneck is therefore well on its way to The expansion of communications services creates new
being overcome. business opportunities and innovative ways of delivering
Domestic fiber-optic networks are needed to carry com- services. The African telecommunications industry itself
munications traffic both within and between countries, has also been a source of business innovation. African
particularly countries that are landlocked. They are there- operators have had to work hard to find ways of rolling out
fore as important as undersea fiber-optic networks for the networks in difficult physical environments and creating

Figure 19.6 Undersea Fiber-Optic Cables in Africa, 2010

IBRD 38556
MAY 2011

London, African Undersea Cables

England

Mediterranean Undersea Cables

Marseille,
France Monaco
Vigo, Spain SEA-ME-WE4, 1280GB, active
Sesimbra,
Portugal Palermo, Italy I-ME-WE, 3840GB, Q4’09
Chipiona, Spain
Asilah, Annaba, Bizerte,
Tripoli, EIG, 3840GB, Q2’10
Tunisia Lebanon
Morocco Algeria
Casablanca, Tripoli,
Morocco Alexandria &
Libya Cairo, Suez, A.R. of Egypt Fujairah,
Altavista, A.R. of United Arab
Canary Is. (Sp) Egypt Emirates Karachi, Pakistan
Jeddah,
Saudi Arabia Oman
Mumbai, India
Nouakchott, Mauritania Port Sudan, Sudan
Dakar, Senegal Massiwa, Eritrea Chennai,
Cape The Gambia India
Verde Guinea-Bissau Abidjan, Côte d’Ivoire Cochin,
Conakry, Djibouti India
Guinea Lomé, Togo
Freetown, Sierra Leone Accra, Cotonou, Benin
Monrovia, Liberia Ghana Lagos, Nigeria
Bonny, Nigeria Colombi, Sri Lanka
Douala, Cameroon Mogadishu,
Bata, Equatorial Guinea Somalia
Libreville, Gabon
São Tomé and Príncipe Mombasa,
Kenya
Pointe-Noire, Congo
Muanda, D.R. of Congo
Cacuaco, Angola Dar es Salaam,
Luanda, Angola Tanzania

Tamatave,
Sub-Saharan Undersea Cables Madagascar
SAT3/SAFE, 380 GB,
active Toliara, Baie du Jacobet,
GLO-1, 640GB, Q4’09 Walvis Bay, Madagascar Mauritius
Maputo, St Paul,
Namibia Mozambique Réunion (Fr)
TEAMs, 1280GB, Q3’09
Seacom, 1280GB, active Mtunzini,
Lion, 1300gb, active South Africa
EASSy, 1400GB, Q2’10 Melkbosstrand,
South Africa
ACE, 1920GB, 2011
MaIN OnE, 1920GB, Q4’10
WACS, 5120GB, Q2’11

Source: http://manypossibilities.net.

344 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES

commercially viable business models based on a customer
base that was once considered marginal. The introduction
of prepaid airtime in very small units is one example of
how African operators have introduced innovations in
their commercial practices to suit the African market.
Zain’s one-network, free regional roaming service is
another innovative business model that originated in
Africa (box 19.2).
Political economy of the sector
The rapid pace of policy reform in the telecommunications
sector in Africa has been influenced by the political econ-
omy of the sector. Three key factors underlie this influence.
First, telecommunications has traditionally not been
available to anyone other than a small elite section of the
population able to afford and obtain a fixed line. But mobile
phones are very popular and there is strong public demand
for them, even in low-income countries. The proven ability
of private companies to develop and operate mobile phone
networks has therefore been less politically controversial
than in other sectors.
Second, the telecommunications business, particularly
mobile telephony, has been very profitable for many
investors. As a result private sector interest in the sector has
grown, and companies have been prepared to pay consider-
able sums for mobile licenses. This has provided an imme-
diate source of revenue for governments that, combined
with ongoing license-fees and tax payments, has made the
telecommunications sector an important source of govern-
ment revenues, which, in turn, has had an influence on gov-
ernment’s willingness to liberalize the market and encourage
private investment.
Box 19.2 Borderless Roaming
In late 2006 Kuwait-based mobile operator Zain (for-
merly Celtel) launched the world’s first borderless
mobile network, One Network, in the Middle East and
East Africa. The service began in 4 countries in East
Africa (Kenya, Sudan, Tanzania, and Uganda) and
expanded to 12 other countries (Burkina Faso, Chad,
the Republic of Congo, the Democratic Republic of
Congo, Gabon, Ghana, Madagascar, Malawi, Niger,
Nigeria, Sierra Leone, and Zambia).
Source: Zain web site (http://www.zain.com/muse/obj/lang.default/portal.view/content/Zainpercent20World/Zainpercent20
Connect/Onepercent20Network) and Business Monitor International (BMI) database (http://www.businessmonitor.com/).
Third, the establishment of global technology standards
in the telecommunications industry and the increasing level
of international competition in the manufacture of
telecommunications equipment have put strong downward
pressure on the cost of building and operating telecommu-
nications networks in Africa. These reduced costs have
allowed many operators to maintain their profitability, even
in the face of increasing competitive pressures.
IMPACT OF ICT IN AFRICA
Several studies examine the macroeconomic impact of ICT.
Roller and Waverman (2001) analyze 21 advanced countries
over 20 years (1970–90). They find a causal nonlinear
relationship between telecommunications infrastructure
investment and economic performance, where impact
increases once countries pass a certain threshold of
telecommunications access. Qiang and Pitt (2004) find that
ICT has made a significant contribution to economic
growth across a wide range of countries.
More recently, research has focused on the economic
impact of broadband Internet. Qiang and Rossotto (2009)
find that every 10 percentage point increase in broadband
Internet penetration in developing countries results in 1.38
percentage points of additional GDP growth. Koutroumpis
(2009) finds a significant and positive relationship between
broadband infrastructure and local, regional, and national
economic growth, including growth in small and medium-
size enterprises. These findings are consistent with other
research that finds that broadband is an important factor in
social transformation and improved service provision, par-
ticularly in rural areas, and that mobile broadband services
are particularly well suited for improving the economic
Through this service, Zain’s customers (both prepaid
and postpaid) can make calls at local rates, roaming
without incurring surcharges. They add air time with
locally purchased airtime cards. In 2009 Zain added
Internet access, e-mail, Multimedia Messaging Service
(MMS), BlackBerry services, Short Messaging Service
(SMS), international roaming, and mobile portal appli-
cations to its One Network local rate, further adding
value to its African customers.
CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 345
 well-being of poor people (Lobo, Novobilski, and Ghosh
2008). Countries that do not develop effective broadband
infrastructure may therefore be failing to capitalize on some
of their economic growth potential (Gaasbeck and Kristin
2008; Tolkoff 2007).
The economic research on the impact of ICT has
focused mainly on the impact at the level of the macro-
economy or firms. ICT is also having a visible impact on
the lives of individuals, changing the way people live and
work in many different ways.
Mobile banking
Banking by cell phone in Africa is one of the most signifi-
cant developments in the recent history of the continent’s
financial sector. The success of some of the early pioneers of
cell phone banking has been replicated in other countries,
through the launch of other types of financial service prod-
ucts delivered by cell phone.
Safaricom’s M-PESA in Kenya was one of the first mobile
banking applications to be launched on the continent (box
19.3 and chapter 20). MAP mobile banking in Uganda fol-
lowed suit. WIZZIT, in South Africa, is another mobile-
based “virtual bank,” whose services can be accessed
through any national mobile phone operator. In addition to
providing a valuable service to customers, WIZZIT has had
a positive economic effect, by employing nearly 2,000 previ-
ously unemployed “WIZZkids” as its sales force. These com-
panies are revolutionizing the financial services sector in
Africa, bringing low-cost financial services to the majority
of the population, that was initially unable to access the tra-
ditional banking sector.
Box 19.3 Serving the Banking Needs of the Poor in Kenya
M-PESA was launched in 2007 to meet the banking
needs of the financially excluded. By July 2010, M-PESA
had more than 11 million customers (about 30 percent
of Kenya’s population) and almost 20,000 agents (up
from 355 at inception). Person-to-person transactions
stood at more than $375 million a month (Safaricom
Ltd, 2010).
There is strong demand for M-PESA’s services in
Kenya, which have had a positive economic impact. Use
of and satisfaction with M-PESA is high. About 40 per-
cent of households use M-PESA (63 percent of them for
regular financial support), 90 percent believe their
346 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES
Telemedicine
Mobile phone operators, health service providers, health
ministries, and donor organizations are working together to
develop innovative ICT-based approaches to healthcare ser-
vice delivery. One such application is TRACnet, which is used
by the Ministry of Health in Rwanda to improve the quality
of service in primary health care institutions (box 19.4).
Agriculture
Agriculture is an essential part of Africa’s economy,
accounting for 13 percent of GDP and employing about 194
million people (World Bank 2010b). ICT is having a positive
impact on the sector in many ways (box 19.5).
IT–enabled services and business process
outsourcing
In addition to being a platform for delivering services, the
ICT sector has the potential to be a source of economic
growth and employment itself. The IT-enabled service
(ITES) sector—also known as business process outsourcing
(ITES-BPO)—is becoming established around the world,
including in some African countries. The sector covers a
wide range of industries from applications development
and services through to IT-enabled services such as call cen-
ters and other types of business process outsourcing.
The ITES-BPO industry is growing and still has consider-
able potential as a source of growth in Africa. Large and small
companies around the world are increasingly hiring compa-
nies in Africa to help them deliver efficient, reliable, and cost-
effective customer support and other key services, such as
money to be safe with M-PESA, 81 percent find it very
easy to use, and 84 percent believed the service to be crit-
ical to their socioeconomic well-being (Agrawal 2010).
Morawczynski and Pickens (2009) find that incomes
of rurual recipients increased 5-30 percent since they
started using M-PESA.
M-PESA has succeeded mainly because it has a
broad market positioning, has a built-in accountability
structure, is easy and safe to use, provides 24/7 support,
is affordable, is provided by the largest mobile phone
network, and has a wide network of agents ensuring
convenience in sending and receiving cash.
 Box 19.4 Improving the Quality of Service Provided by Primary Health Care Facilities in Rwanda

TRACnet is a mobile phone–based platform for moni-
toring HIV treatment in Rwanda. By 2009 it had regis-
tered more than 1,000 service providers, conducted
more than 85,000 annual user sessions, and collected
longitudinal data on more than 105,000 patients.
Thanks to TRACnet, Rwanda has access to robust
datasets of HIV/AIDS patients located centrally and
accessible from any location, allowing faster and better-
informed intervention. As a result, public monitoring of
HIV/AIDS transmission patterns has improved. Doc-
tors and patients also have instantaneous access to more
reliable information. Real-time monitoring of anti-
retroviral drug stocks leads to quicker replenishments.
Source: http://www.un.org/esa/sustdev/publications/africa_casestudies/tracnet.pdf; http://www.kiwanja.net/database/project/
project_voxiva_hivaidsrelief.pdf.
The improved information exchange between remote
health facilities and central actors has reinforced
accountability in care and treatment of patients.
TRACnet has succeeded because it is based on
simple technology using the widely available platform
of the mobile phone. The centralized database is
interoperable with multiple communication channels
and it uses open source software giving it flexibility
and scalability. It has the full support of the Rwanda
government—a key to its success—and the program
includes a training component that ensures that
health workers are well prepared to work with the
system.

Box 19.5 Increasing Access to Market Information in West Africa

Esoko leverages mobile phones to enhance productivity Transactions costs for farmers and traders have also
gains for African farmers and traders by giving them fallen by $2–$150 per transaction by significantly
quicker access to better market information. Starting in reducing the role of middlemen or cutting them out
2007 in Ghana with only $90,000 and operating in the altogether. It has also transformed mobile phones into a
red by up to $21,000, Esoko had broken even by the market bulletin increasing their utility beyond voice
fourth year and by the fifth year had $1.4 million in rev- and text.
enues and about $540,000 in profits. By 2010 Esoko had Esoko has succeeded mainly because it uses open
been scaled to seven other countries (Benin, Burkina source software, enabling it to scale up, tailor business
Faso, Cameroon, Côte d’Ivoire, Madagascar, Mali, and services to local needs, use affordable mobile telephony,
Togo) and had grown to 40 full-time employees and offer free listing of services, allow sending of and receipt
about 9,000 users. of text messages in several languages, provide real-time
Esoko has led to a 6.4 percent fall in grain price mar- commodity prices, and provide direct access to markets
ket dispersion and a 3.5 percent decline in mean prices. worldwide.
Source: http://www.esoko.com/; http://www.slideshare.net/slavb/s-bartlett-esoko-cirad-2010; http://www.ifc.org/ifcext/
spiwebsite1.nsf/f451ebbe34a9a8ca85256a550073ff10/0e6f5be010f90329852576ba000e2dad?OpenDocument; http://www.slideshare
.net/slavb/davies-esoko-cirad-2010.

data entry and document processing. A 2010 study by
McKinsey and Company estimates that the addressable mar-
ket for IT and ITES offshoring is $500 billion a year, of which
only about 20 percent has been realized (Sudan et al. 2010).
Success in the ITES–BPO sector brings with it a number
of benefits, including the following:
■ Employment of women. Women now account for a large
percentage of all professional and technical workers
in the IT/ITES sector—a much higher rate of female
participation than in the service sector in general
(Sudan et al. 2010).
■ Increase in investment. The IT/ITES sector helps attract
foreign investment, transform the financial sector, ener-
gize local exports, and nurture ICT skills and innovation
in the workforce.
■ Job creation. It is estimated that every job created in the
IT/ITES sector results in the creation of four additional
jobs in ancillary sectors, such as transport, training, and
catering.

CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 347

 Ghana and South Africa have identified ITES as one of
the key sectors for enhancing economic growth. Both seek
to position themselves as premier BPO destinations in
Africa (box 19.6).
THE ROLE OF THE WORLD BANK GROUP
The World Bank Group has been involved in the ICT sec-
tor in Africa since 1969, when the International Develop-
ment Association funded an $800,000 telecommunication
project in Burkina Faso. Its work has covered investment
lending, policy and regulatory reforms, privatization,
and e-government and applications projects. The Bank
has provided technical assistance to 27 Sub-Saharan
African governments to strengthen policy and regulatory
frameworks and build institutional capacity in the
sector, including capacity for regulatory design and
implementation.
In 2007 the World Bank Group launched an initiative to
enhance broadband connectivity in Eastern and Southern
Africa through the Regional Communications Infrastruc-
ture Program (RCIP). The program is currently operational
in seven countries (Burundi, Kenya, Madagascar, Malawi,
Mozambique, Rwanda, and Tanzania) and may be expanded
to other countries in the region. Similar projects are under
Box 19.6 Expanding Business Process Outsourcing in Ghana and South Africa
In Ghana the government’s proactive policies of sector
reform has created a competitive telecommunications
industry with a telephone penetration of more than
60 percent. This industry is providing a platform on
which an IT and IT-based services industry is growing.
One of the efforts, an $84.4 million World Bank
project, is supporting is the development of a busi-
ness process outsourcing center. More than 1,000 jobs
were created in the industry in 2009 and 2010. Ghana
estimates that it will create some 37,000 jobs by 2011,
increasing the sector’s contribution to GDP by about
$750 million. The 2009 AT Kearney Global Services
Location Index ranks Ghana 1st out of 50 countries in
the world in terms of financial attractiveness and 15th
in terms of location attractiveness. The partnership
between the public and private sector to train some
50 training providers and 6,000 business process
Source: http://www.busrep.co.za/index.php?fSectionId=561andfArticleId=5554024 http://retail.bizcommunity.com/Article/196/458/
49992.html.
Notes: (http://www.ites.gov.gh/IT-Industry.aspx).
348 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES
development in Central Africa (Cameroon, the Central
African Republic, and Chad) and West Africa.
The Bank is also supporting governments in their efforts
to use the ICT infrastructure to improve service delivery.
In three stand-alone projects (e-Benin, e-Ghana, and
e-Rwanda), it is providing funds to incorporate ICT into the
delivery of public services. Implementation of the RCIP
program in Kenya, Mozambique, and Tanzania also has
major e-government components. ICT is also integrated
throughout the broader World Bank portfolio of projects
and is a major component of the Bank’s research, dissemi-
nation, and policy development work in Africa. This work
has included major outreach events, such as an innovations
conference day held as a part of the Africa Union Summit in
Ethiopia in 2010 and a seminar on “The Transformational
Power of ICT for Africa” at the 2010 World Bank Group
Spring Meetings.
The World Bank is also supporting African govern-
ments in their efforts to use the growing ICT infrastruc-
ture to develop the IT and ITES-BOP industry. Many of
the ICT sector projects in Africa include components that
support the industry through training, regulatory reforms,
and, in some cases, investment in supporting infrastruc-
ture. A related initiative is the New Economy Skills for
Africa Program (NESAP), a joint effort by World Bank
outsourcing agents is expected to further strengthen
Ghana’s position as one of the most attractive loca-
tions for ITES business in Africa.
In July 2010 Amazon.com, the world’s largest Internet
retailer, announced that it would open a new customer
service call center in Cape Town, South Africa, where it
will operate a software development center. According to
the local media, this business development was brought
about by the strong commitment by the central and
Western Cape governments for the development of the
ITES-BPO industry. It is estimated that the call center
will create 1,000 new jobs (600 permanent positions and
400 seasonal positions). Servicing will be provided in
English and German. Local government officials hope
that the deal will boost the local economy not only by
creating jobs but also by demonstrating that Cape Town
has the capacity to host world-class clients.
 Box 19.7 Supporting Skills Development in Africa through the New Economy Skills for Africa
Program–Information and Communication Technologies (NESAP–ICT) Initiative
NESAP–ICT is the World Bank’s cross-sectoral initia-
tive to support the development of ICT skills.
Launched in 2008 in eight Sub-Saharan African
countries (Ghana, Kenya, Madagascar, Mozambique,
Nigeria, Rwanda, Senegal, and Tanzania), the pro-
gram seeks to bridge education and industry gaps
in the field, create a globally benchmarked assess-
ment talent pool for the ITES-BPO industry, and
strengthen collaboration with IT industry leaders and
associations.
The initiative consists of two focus areas. Window I
focuses on developing ICT skills for a new emerging
ICT-based sector. Window II focuses on ICT use in
education.
education, private sector development, and ICT sector
units to provide skills development for the ICT industry
(box 19.7).
In parallel with World Bank activities, IFC has been
closely involved in the sector since the beginning of the
mobile revolution in Africa. It committed a $261 million
Implementation of Window I began in 2008. The
various skills required for the IT/ITES sector were iden-
tified and segmented to meet each country’s needs.
Fifty-four delegates from eight countries participated in
a study tour of India. Pilot projects, guided by interna-
tional best practice, were launched in four countries
(Ghana, Kenya, Nigeria, and Tanzania). The program
established partnerships with the world’s leading ICT
companies (including Microsoft, Intel, Cisco, Oracle,
IBM, Nokia, and EMC); with learning institutions (for
example, Carnegie Mellon University); and with IT-
BPO industry associations (for example, NASSCOM in
India). Implementation of Window II, expected to begin
in 2011, will build on the Window I accomplishments.
investment to 10 ICT projects in Africa during fiscal 2010. It
is also investing in innovative new ICT businesses, such as
WIZZIT in South Africa, which is providing mobile-based
banking services, and Helios Towers Africa Ltd., which is
providing tower infrastructure services to mobile operators
in Nigeria.
CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 349
350

Table 19A.1 Privatizations of Telecom Incumbents in Sub-Saharan Africa, 1995–2010

Percent
Initial privatization transaction
Percent Amount private
Country Operator Date sold ($millions) 2008 Note
Burkina Faso ONATEL December 2006 51 295 51 Private sale to Maroc Telecom.
Cape Verde Cabo Verde Telecom December 1995 40 20 59 Private sale to Portugal Telecom. Subsequent distribution to
employees (5 percent of total), national private investors
(14 percent), and government social security system
(38 percent).
Central African Rep. Socatel — — — — France Cable and Radio owned 40 percent of shares at one
point. Current status not available.
Côte d’Ivoire Côte d’Ivoire Telecom January 1997 51 210 51 Private sale to France Telecom.
Equatorial Guinea Getesa 1987 40 40 Private sale to France Telecom.
Gabon Gabon Telecom February 2007 51 79 51 Private sale to Maroc Telecom.
Gambia, The GAMTEL January 2007 50 35 50 Private sale to Spectrum Investment Holding (Lebanon).
Ghana Ghana Telecom December 1996 30 38 70 Original private sale to G-Com consortium headed by
Telekom Malaysia. In 2002 the government of Ghana
abrogated the management contract with G-Com and
bought back shares. Subsequent private sale of 70 percent
to Vodafone (UK) in August 2008 for $900 million.
Guinea Sotelgui December 1995 60 45 0 Renationalized in 2008 following private sale to Telekom
Malaysia.
Guinea-Bissau Guinée Telecom 1989 51 3 0 Renationalized following private sale to Marconi (later
assumed by Portugal Telecom).
Kenya Telkom Kenya December 2007 51 390 51 Sale to consortium led by France Telecom (78.5 percent)
with Alcazar Capital Ltd. (21.5 percent).
Lesotho Telecom Lesotho November 2000 70 — 70 Private sale to Mountain Communications (Econet)
(Zimbabwe), Mauritius Telecom, and Eskom (South Africa).
Sale price not disclosed.
Madagascar TELMA August 2003 34 12.6 68 Private sale to Distacom (Hong Kong, China), which also
purchased France Telecom’s ownership.
Malawi Malawi Telecom (MTL) February 2006 80 30 80 Private sale to Telecom Holdings Ltd. (THL) consisting of
PCL (50.1 percent), Old Mutual (16.1 percent), NICO
(5.0 percent), Detecon (Germany) (2.6 percent) and
Press Trust (6.2 percent). Percentages refer to MTL
ownership.
Mali SOTELMA July 2009 51 384 51 Private sale to Maroc Telecom.
Mauritania Mauritel April 2001 54 48 54 Private sale to Maroc Telecom, which subsequently engaged
in a series of sales with local investors. Its ownership
stood at 51 percent in 2008.
Mauritius Mauritius Telecom November 2000 40 261 40 Private sale to France Telecom.
Niger SONITEL November 2001 51 16 51 Private sale to Chinese and Libyan consortium. Government
has announced intention to renationalize.
 Nigeria NITEL July 2006 51 500 Private sale to TransCorp (Nigeria). Government has
rescinded the sale and was in process of reprivatizing
in 2010.
Rwanda Rwandatel June 2005 99 20 80 Initial private sale to Terracom (United States), which
government later repurchased. Libya Arab Portfolio later
acquired an 80 percent interest for $100 million.
São Tomé and Principe CST (Companhia Santomense de 1989 51 1 51 Private sale to Portugal Telecom.
Telecomunicações)

Senegal Sonatel July 1997 33 90 73 Initial private sale to France Telecom. Subsequent additional
sale to France Telecom and listing on regional stock
exchanges.
Seychelles Cable and Wireless Seychelles 1981 49 n.a. 100 Government granted the British company Cable and
Wireless the right to operate the telephone network,
inaugurated in 1954. Cable and Wireless was
subsequently privatized by British government in three
tranches (1981, 1983, and 1985).
South Africa Telkom May 1997 30 1,261 34 Initial private sale to Thintana Communications
(SBC [United States] 60 percent) and Telekom Malaysia
(40 percent). Global initial public offering in March 2003
of 47 percent. Thintana Communications sold
14.9 percent interest in Telkom to South African and
international institutional investors in June 2004 and its
remaining interest to the Public Investment Corporation,
wholly owned by the South African Government in
November 2004. Subsequent Telkom share repurchases
have altered its level of private shareholding.
Sudan Sudatel 1997 n.a. n.a. 36 Multiple share offerings on local and regional stock
exchanges.
Tanzania TTCL (Tanzania Telecommunica- February 2001 35 65 35 Private sale to a consortium of MSI (Netherlands) and
tions Company Ltd.) Detecon (Germany).
Uganda Uganda Telecom May 2000 51 34 69 Initial private sale to UCOM consortium consisting of
Detecon (Germany), Telecel (Switzerland), and Orascom
(Egypt). UCOM ownership subsequently purchased by
Libyan Arab Portfolio (LAP), which then increased its
ownership through a capital increase.
Zambia Zamtel June 2010 75 394 n.a. Privatization of 75 percent stake sold to LAP Green (Libya)
for $394 million in 2010.

Source: AICD 2009; Business Monitor International (BMI) database (http://www.businessmonitor.com/).

Note: n.a. = Not applicable. — = Not available.
351
 REFERENCES
Agrawal, M. 2010. “M-Pesa: Transforming Millions of Lives – I.”
http://www.telecomcircle.com/2010/01/m-pesa/.
AICD (Africa Infrastructure Country Diagnostic). 2009.
Africa Infrastructure Country Diagnostic Background
Paper 10. Information and Communications Technology in
Sub-Saharan Africa: A Sector Review. Washington, DC.
———. 2010. Africa’s Infrastructure: A Time for Transforma-
tion. Washington, DC.
———. Forthcoming. ICT Sector Volume. Washington, DC.
Gaasbeck, V., and A. Kristin. 2008. “A Rising Tide: Measuring
the Economic Effects of Broadband Use across California.”
Social Science Journal 45 (4): 691–99.
International Telecommunications Union (ITU). 2010.
World Telecommunication/ ICT indicators Database.
Geneva, Switzerland.
Koutroumpis, P. 2009. “The Economic Impact of Broadband
on Growth: A Simultaneous Approach.” Telecommunica-
tions Policy 33 (9): 471–85.
Lobo, B. J., A. Novobilski, and S. Ghosh. 2008. “The Economic
Impact of Broadband: Estimates from a Regional Input-
Output Model.” Journal of Applied Business Research 24
(2): 103–14.
Morawczynski, O., and M. Pickens. 2009. Poor People Using
Mobile Financial Services: Observations on Customer Usage
and Impact from M-PESA. Consultative Group to Assist
the Poor (CGAP), Washington, DC.
352 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES
Qiang, C. Z-W., and A. Pitt. 2004. “Contribution of Infor-
mation and Communication Technologies to Growth.”
World Bank Working Paper 24, Washington, DC.
Qiang, C. Z-W., and C. M. Rossotto, with K. Kimura. 2009.
“Economic Impact of Broadband.” In Information and
Communication for Development: Extending Reach and
Increasing Impact, 35–50. Washington, DC: World Bank.
Roller, L-H., and L. Waverman. 2001. “Telecommunica-
tions Infrastructure and Economic Development: A
Simultaneous Approach.” American Economic Review
91 (4): 909–23.
Safaricom Ltd. 2010. http://www.safaricom.co.ke/
Sudan, R., S. Ayers, P. Dongier, A. Muente-Kunigami, and
C. Zhen-Wei Qiang. 2010. The Global Opportunity in
IT–Based Services: Assessing and Enhancing Country
Competitiveness. Washington, DC: World Bank.
TeleGeography CommsUpdate. 2009. July 8. http://www
.telegeography.com/cu/index.php.
Tolkoff, S. 2007. “Increased Broadband Use Could Add
186,000 Jobs.” Orange County Business Journal 30 (50): 9.
Williams, Mark D. J. 2010. Broadband for Africa: Developing
Backbone Communications Networks. Washington, DC:
World Bank.
World Bank. 2010a. Public Private Infrastructure Indicators
Database. Washington, DC.
———. 2010b. World Development Indicators. Washington,
DC.
 CHAPTER 20
Mobile Payments Go Viral
M-PESA in Kenya
Ignacio Mas and Dan Radcliffe
-PESA is a small-value electronic payment and
M
store-of-value system in Kenya accessible from
ordinary mobile phones. It has seen exceptional
growth since its introduction in March 2007. Now in
use by 9 million customers—40 percent of Kenya’s adult
population—the system processes more transactions
domestically than Western Union does globally. M-PESA’s
market success is the result of the interplay of three fac-
tors: preexisting country conditions that made Kenya a
conducive environment for a successful mobile money
deployment; a clever service design that facilitated rapid
adoption and early capturing of network effects; and a
business execution strategy that helped M-PESA rapidly
reach a critical mass of customers, thereby avoiding the
adverse chicken-and-egg (two-sided market) problems
that afflict new payment systems.
M-PESA IN A NUTSHELL
M-PESA (“M” for mobile and “PESA” for money in
Swahili)was developed by mobile phone operator Vodafone
and launched commercially by its Kenyan affiliate Safari-
com in March 2007.1 To access the service, customers must
first register at an authorized M-PESA retail outlet. They are
then assigned an individual electronic money account
linked to their phone number and accessible through an
application stored on the subscriber identification module
(SIM) cards of their mobile phones.2 The application has
two main functions. First, it allows customers to deposit
cash to and withdraw cash from their accounts by exchang-
ing cash for electronic value at a network of retail stores.
Second, it allows users to transfer funds to others, to pay
bills, and to purchase mobile airtime credit. Retail stores are
paid a fee by Safaricom each time they exchange cash for M-
PESA credit on behalf of customers. All M-PESA transac-
tions are authorized and recorded in real time using secure
short messaging service (SMS) and are capped at $500.
M-PESA registration is free, as is making deposits into
the system. Customers are charged flat fees of approximately
$0.403 for person-to-person (P2P) transfers and bill pay-
ments, $0.33 for withdrawals (for transactions under $33)
and $0.013 for balance inquiries. Individual customer
accounts are maintained in a server that is owned and man-
aged by Vodafone, but Safaricom deposits the full value of
its customers’ balances in the system in pooled accounts in
two regulated banks. Thus, while Safaricom issues and man-
ages the M-PESA accounts, the value of the accounts is fully
backed by highly liquid deposits at commercial banks.
Rather than paying customers interest on the balance in
their M-PESA accounts, Safaricom sets aside a small per-
centage of account balances in a not-for-profit trust fund.
The purpose of these funds has not yet been decided.
M-PESA’s function as a retail payment platform is
important because it reaches a large number of people com-
pared with other financial services outlets in Kenya (figure
20.1).4 There are now nearly five times as many M-PESA
outlets in Kenya as there are PostBank branches, post
offices, bank branches, and automated teller machines
353
 Figure 20.1 Outlets Offering Financial Services in Kenya M-PESA trials. DFID’s role in spotlighting the need for
mobile payments and funding the early risk demonstrates
100,000 good practice for donor funding.
100,000

16,900
Number of outlets

A snapshot of M-PESA after three years

10,000
The speed and extent to which M-PESA has been deployed
1,510 in Kenya is remarkable (figure 20.2). The number of cus-
1,000 800 840
440 tomers hit the 9 million mark in November 2009, less than
three years after the service was launched. This number of
100
customers represents 60 percent of Safaricom’s customer
base, 40 percent of Kenyan adults, and 23 percent of the
cie s

Ba fice t

es

rs
of pos
en he

re

lle
ch

AT
To s

to
ag c

country’s total population.5

se
d ran

an
l

as
ta

re
br
an k b

es

e
nk

-P By other measures, too, M-PESA has a deep reach in

im
n
ba

rt
st

Ai
Kenya.6 The number of retail stores at which M-PESA users
Po

Sources: Central Bank of Kenya, Kenya Post Office Savings Bank, and
Safaricom.
(ATMs) combined. M-PESA’s presence in rural areas is par-
ticularly important, because access to financial services in
such areas is limited, and the ability to use existing retail
stores as M-PESA cash-in/cash-out outlets reduces deploy-
ment costs, provides greater convenience, and lowers the
cost of access compared with other financial services outlets.
Importantly, both private and public actors were
involved in creating and enabling M-PESA. The idea of
M-PESA was conceived by a London-based team within
Vodafone. This team believed that mobile phones could play
a central role in lowering the cost of access to financial ser-
vices for poor people. The idea was then developed by the
Safaricom team in Kenya, which customized it and oversaw
a very focused execution. The Central Bank of Kenya, in par-
ticular the Payments System Group, helped to enable the
launch of M-PESA by allowing a mobile operator to take the
lead in providing payment services to the general population.
Following the first FinAccess survey in 2006, which showed
very low levels of bank penetration in Kenya, the central bank
was determined to explore all reasonable options for correct-
ing the financial access imbalance. It worked in close partner-
ship with Vodafone and Safaricom to assess the opportunities
and risks involved prior to the launch of M-PESA. Conscious
that premature regulation could stifle innovation, the Central
Bank of Kenya chose to closely monitor and learn from early
M-PESA trials and to formalize regulations later.
Finally, the United Kingdom’s Department for Interna-
tional Development (DFID) played an instrumental role in
the creation of M-PESA within Kenya, first by funding the
organizations that made the FinAccess survey possible, and
then by providing seed funding to Vodafone for early
can cash in and cash out now totals 16,900, of which nearly
half are located outside urban centers. M-PESA now han-
dles $320 million in P2P transfers per month. On an annu-
alized basis, this is equal to roughly 10 percent of Kenya’s
gross domestic product (GDP).7 An average of $650 million
in cash deposit and withdrawal transactions is made at
M-PESA stores every month, with an average transaction
size of approximately $33. Nearly one-fifth of Safaricom air-
time purchases are now conducted through M-PESA.
M-PESA’s bill pay function, launched in March 2009, has
been popular: 75 companies now use M-PESA to collect pay-
ments from their customers. The biggest user, the electric
utility company, reports that roughly 20 percent of its 1 mil-
lion customers now pay through M-PESA. At least two banks,
Family Bank and Kenya Commercial Bank, are now using
M-PESA as a mechanism for customers to either repay loans
or withdraw funds from their banks accounts. And 27 com-
panies are using M-PESA for bulk payment distribution.
Customer perspectives on M-PESA
A survey of 3,000 M-PESA users and nonusers conducted in
late 2008 shed considerable light on the profile of M-PESA’s
early adopters and customer usage patterns. Compared with
nonusers, the average M-PESA user is twice as likely to have
a bank account (72 percent versus 36 percent) and is wealth-
ier (65 percent higher expenditure levels), more literate, and
better educated (Suri and Jack 2008). Early adopters of the
service also appear to be experienced with banking services
and fairly savvy with technology, which probably makes
them more keenly aware of the convenience of M-PESA rel-
ative to alternative financial services.
Figure 20.3 highlights the ways customers report they use
M-PESA. More than half the sample use the service primar-
ily for sending and receiving money, a use consistent with

354 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA

Figure 20.2 Growth of the M-PESA Customer Base

10

Number of customers (in millions)

9
8
7
6
5
4
3
2
1
0
7

07

08

09

9
l-0

-0

-0

-0

-0

l-0

-0

-0

-0

-0

l-0

-0
p-

p-

p-
ov

Jan

ar

ay

ov

Jan

ar

ay

ov
Ju

Ju

Ju
Se

Se

Se
M

M
M

M
N

N
Source: Safaricom.

Figure 20.3 How M-PESA Customers Use the Service Ninety-eight percent of M-PESA users are happy with
the service, according to the 2008 survey, and 84 percent
Pay bills, 2% claim that losing M-PESA would have a large, negative effect
Store/save money
for emergencies, 14% None of these, 1% on them (Suri and Jack 2008). Figure 20.4 illustrates how
customers compare M-PESA with alternative services. In all
Buy airtime for categories—speed, convenience, cost, and safety—customers
someone else, reported better service with M-PESA than with other forms
8%
of financial services.
Receive money,
28% M-PESA’s service evolution

M-PESA’s original core offering, the P2P payment, enabled

Buy airtime for
myself, 14% customers to send money to anyone with access to a mobile
phone. Importantly, it opened up a market for transactions
that previously were handled largely informally—through
Store/save
money for
Send money, personal trips, friends, and public transport networks (“per-
25%
everyday use, sonal networks” in figure 20.5). Although many transactions
14% carried out under M-PESA, such as sending a portion of
salary earned at the end of the month to relatives, can be
characterized as scheduled payments, others allow people
Source: Suri and Jack 2008. to draw on a much broader network of family members,
friends, and business associates to access money when
required. Thus, in addition to providing a large measure of
M-PESA’s broad market positioning. Though 21 percent convenience for transactions that were already occurring,
report using M-PESA for storing money, the survey revealed M-PESA also provides a basic form of financial protection
that less than 1 percent of accounts had balances of more for a large number of users by enabling a network for
than K Sh 1,000 ($13), and a government audit (Okoth instant, on-demand payments.
2009) of M-PESA in January 2009 showed that the average In recent months, Safaricom has increasingly opened up
balance on M-PESA accounts was only $2.70. The survey M-PESA to institutional payments, enabling companies to
also found that 52 percent of customers use the service on pay salaries and collect bill payments. In the future, Safari-
only a monthly basis, suggesting that customers have yet com envisions increased use of M-PESA for e-commerce
to incorporate M-PESA into their daily lives (Suri and and in-store purchases, a strategy represented by the down-
Jack 2008). ward arrow in figure 20.5.

CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 355

 Figure 20.4 User Ratings of M-PESA Compared with Alternatives

Slower Less
2% convenient
4%

Speed Convenience
More
Quicker More
convenient
98% Less safe expensive
96%
2% 4%

Safety Cost

Safer Cheaper
98% 96%

Source: Suri and Jack 2008.

Note: At the time of the survey, alternatives consisted of 876 bank branches, 1,025 Post Office branches, 1,424 ATMs and 6,104 M-PESA agents.

Figure 20.5 Potential Range of Transactions Supported by M-PESA

FORMAL FINANCIAL
M-PESA role PRODUCTS
in promoting savings, credit, insurance
fuller financial
inclusion?
INFORMAL SERVICE
PROVIDERS
pawnbroker, moneylender
PUSHING &
PULLING MONEY
ACROSS TIME “On-demand” payments
PERSONAL NETWORKS
“JUST “scheduled” payments
PAYMENTS”

REMOTE B2C/C2B
INSTITUTIONAL PAYMENTS
M-PESA as salaries, bill pay, G2P,
a fuller retail online/e-commerce
payments
platform
IN-STORE MERCHANT PAYMENTS
for goods & services

Source: Bill & Melinda Gates Foundation analysis.

As represented by the upward arrow in figure 20.5,
another of M-PESA’s goals is to become a vehicle for delivery
of a broader range of financial services for the Kenyan pop-
ulation. Thus far, evidence that people are willing to use the
basic M-PESA account as a store of value is limited. There is
likely, however, to be a need to develop more targeted savings
products that balance customers’ preference for liquidity
and commitment and that will connect to a broader range
of financial institutions. This is the path M-PESA must take
to deliver on its promise of addressing the challenge of
financial inclusion in Kenya. A key precondition is regula-
tion: the Central Bank of Kenya is in the process of finalizing

356 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA

regulations that will allow nonbank outlets and platforms
such as M-PESA become channels for formal deposit-taking.
Beyond that, Safaricom will need to develop appropriate
service, commercial, and technical models under which
M-PESA can interface with the systems of other financial
service providers.
The broader lessons from M-PESA’s success
In addition to the compelling marketing, cold business
logic, and consistent execution behind M-PESA, the success
of the service is also a vivid example of how great things can
happen when public and private organizations rally around
common challenges and innovative solutions. Three top-
line lessons emerge from M-PESA’s success. First, M-PESA
shows the promise of leveraging mobile technology to
extend financial services to a large number of unbanked
poor people. Second, it demonstrates the importance of
designing usage-based rather than float-based revenue
models for financial services for poor customers. And third,
M-PESA demonstrates the importance of building a low-
cost transactional platform that enables customers to meet
a broad range of payment needs.8
Leveraging technology to extend financial services
to unbanked poor people. Mobile phone technology is
quickly becoming ubiquitous, even among poor segments of
the Kenyan population. Mobile penetration in Africa has
increased from 3 percent in 2002 to 48 percent in 2010, and
is expected to reach 72 percent by 2014, according to Wire-
less Intelligence. The mobile device mimics some of the key
ingredients needed to offer banking services. The SIM card
inside GSM phones, for example, can be used to authenti-
cate users, thereby avoiding the high costs of distributing
separate bank cards to poor customers, who typically do not
generate significant profits for banks. The mobile phone can
also be used as a point of sale terminal to initiate financial
transactions and securely communicate with the appropriate
server to request transaction authorization, thus obviating the
need to deploy costly dedicated devices in retail environments.
A usage-based model for reaching poor customers
with financial services. Because banks make most of
their profits by collecting and reinvesting deposits, they tend
to distinguish between profitable and unprofitable cus-
tomers based on the likely size of their account balances and
their ability to absorb credit. Banks thus find it difficult to
serve poor customers because the revenue from reinvest-
ing small-value deposits is unlikely to offset the cost of
serving these customers. In contrast, mobile operators in
developing countries have developed a usage-based revenue
model, selling prepaid airtime to poor customers in small
increments, so that each transaction is profitable on a stand-
alone basis. This is the magic behind the rapid penetration
of prepaid airtime into low-income markets: a card pur-
chased is profit booked, regardless of who buys the prepaid
card. This usage-based revenue model is directly aligned
with the model needed to sustainably offer small-value
cash-in/cash-out transactions at retail outlets and would
make possible a true mass-market approach, with no incen-
tive for providers to deny service based on minimum bal-
ances or intensity of use.
A low-cost transactional platform that enables
customers to meet a range of payment needs. Once
customers are connected to an e-payment system, they can
use this capability to store money in a savings account, send
and receive money from friends and family, pay bills and
monthly insurance premiums, receive pension or social wel-
fare payments, or receive loan disbursements and repay
them electronically. In short, when customers are connected
to an e-payment system, their range of financial possibilities
expands dramatically.
Putting these elements together, M-PESA has prompted
a rethink on the optimal sequencing of financial inclusion
strategies. Whereas most financial inclusion models have
employed “credit-led” or “savings-led” approaches, the
M-PESA experience suggests that there may be a third
approach—focusing first on building the payment “rails” on
which a broader set of financial services can ride.
KENYA COUNTRY FACTORS: UNMET NEEDS,
FAVORABLE MARKET CONDITIONS
The growth of M-PESA is a testament to Safaricom’s vision
and execution capacity. But Safaricom also benefited from
launching the service under several enabling conditions for
successful deployment of a mobile money service. These
include strong latent demand for domestic remittances,
poor quality of available financial services, a banking regu-
lator that permitted Safaricom to experiment with different
business models and distribution channels, a mobile com-
munications market characterized by Safaricom’s dominant
market position and low commissions on airtime sales, and
a reasonable amount of banking infrastructure.
Strong latent demand for domestic remittances
Safaricom based its launch of the M-PESA service on the
very brief, but powerful, phrase: “send money home.” In
CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 357
 this, it capitalized on the fact that demand for domestic
remittance services is greater in locations where migration
has occurred, separating families, particularly when the
breadwinner moves to an urban center and the rest of the
family remains home. This was certainly the case in Kenya,
where 17 percent of households depended on remittances as
their primary income source as of 2006 (FSDT 2007a).
A recent study of M-PESA suggests that latent demand
for domestic remittances is related to urbanization ratios
(Ratan 2008), and that the most propitious domestic remit-
tances markets are those in which the process of rural-
urban migration is sufficiently rooted to produce large
migration flows but not so advanced that rural communi-
ties are hollowed out. The study also finds that countries
with mid-range urbanization ratios (20 to 40 percent), espe-
cially those that are urbanizing at a rapid rate, are likely to
exhibit strong rural-urban ties requiring substantial transfer
of value. This is the case in many African countries, such as
Kenya and Tanzania, where the share of the population liv-
ing in urban areas as of 2008 was 22 percent and 26 percent,
respectively, according to the World Bank. On the other
hand, where urbanization ratios exceed 50 percent, such as
in the Philippines and several Latin American countries,
remittances are likely to be more closely linked to interna-
tional rather than domestic migration patterns.
Further, the study shows that in locations where entire
nuclear families move, remittances are strongest when there
is cultural pressure to retain a connection with one’s ances-
tral village. In Kenya, migrants’ ties with rural homes are
reinforced by an ethnic (rather than a national) concept of
citizenship. These links are expressed through burial, inher-
itance, cross-generational, social insurance, and other ties,
even in cases where migrants reside more or less perma-
nently in cities.9 In countries where migrants have a
stronger connection to national as opposed to local or eth-
nic identity, rural to urban migration may have diminished
the significance of the rural “home” and hence dampened
domestic remittance flows.
Poor quality of existing alternatives
Demand for mobile e-payments must be examined in the
context of the accessibility and quality of alternative pay-
ment methods. If there are many good alternatives to
mobile payments, as is typically the case in developed coun-
tries, it is difficult to convince users to switch to the new ser-
vice. In the Philippines, for example, the G-Cash and Smart
Money mobile payment services experienced low take-up in
part because of the availability of a competitive alternative
358 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA
to mobile payments—an extensive and efficient semiformal
retail network of pawnshops offering domestic remittance
services at commissions of 3 percent.
In Kenya, by comparison, the most common channel for
sending money before M-PESA was informal bus and
matatu (shared taxi) companies. Because these companies
are not licensed to transfer money, there is considerable risk
that the money will not reach its final destination. Mean-
while, Kenya Post, Kenya’s major formal remittance
provider, is perceived by customers as costly, slow, and
prone to liquidity shortages at rural outlets. M-PESA’s pop-
ularity was also bolstered by the fact that Kenya’s bank
branch infrastructure (currently, there are 840 branches) is
far too sparse to compete with M-PESA’s 16,900 cash-
in/cash-out outlets. Figure 20.6 illustrates how Kenyan
households sent money before and after M-PESA. Of note is
the dramatic reduction in the use of informal bus systems
and Kenya Post to transfer money between 2006 and 2009.
As noted, M-PESA’s early adopters were primarily
banked customers, suggesting that M-PESA did not acquire
its initial critical mass through competition with the formal
sector but rather as a complement to formal services for
clients who were wealthier, more exposed to formal finan-
cial service options, and less risk averse. As M-PESA services
move deeper into the Kenyan market, however, unbanked
users are increasingly driving M-PESA’s expansion because
of the competitive advantages of mobile banking offers over
other options. This is one reason why Africa, with its high
proportion of unbanked people, is seen as such a promising
market for mobile money applications.
A supportive banking regulator
Regulation of mobile money can help secure trust in new
mobile money schemes. At the same time, regulation may
constrain the deployment of a mobile money application by
limiting the scheme operator’s freedom in structuring the
business model, service proposition, and distribution chan-
nels. In the case of M-PESA in Kenya, Safaricom had a good
working relationship with the central bank and was given
regulatory space to design M-PESA in a manner that fit its
market. Together, the Central Bank of Kenya and Safaricom
worked out a model that provided sufficient prudential
comfort to the central bank.
The Central Bank of Kenya insisted that all customer
funds be deposited in a regulated financial institution and
reviewed the security features of the technology platform,
but it also allowed Safaricom to operate M-PESA as a pay-
ments system outside the provisions of the banking law.10
Figure 20.6 Money Transfer Behavior before and after M-PESA

50
45
40
35
30
Percent 25
20
15
10
5
0
A

ce

sit

er
Bu

sfe

qu

els
an
ES

th
fi

po
of
H

he
an
-P

O
ne
de

tr
M

C
st

eo
Po

ct

ey

m
ire

on

So
D

M
2009 2006

Source: FSDT 2007a, 2009b.

Safaricom has paid a certain price for this arrangement. For
instance, interest earned on deposited balances must go to a
not-for-profit trust and cannot be appropriated by Safari-
com or passed on to customers. To address anti-money-
laundering concerns, there are also limits on the size of
M-PESA transactions. Fundamentally, however, Safaricom
was able to design the M-PESA service without having to
contort its business model to fit within a prescribed regula-
tory model.
The Central Bank of Kenya has continued to support
M-PESA’s development, even in the face of pressure from
banks. In late 2008, following a lobbying effort by the bank-
ing industry to shut down the service, the Central Bank of
Kenya performed an audit of the M-PESA service at the
request of the Ministry of Finance and declared it safe and
in line with the country’s objectives for financial inclusion
(see Okoth 2009 for information). Thus far, the central bank
appears justified in its confidence in M-PESA—there have
been no reports of major fraud. Although system downtime
remains frequent, it has not been catastrophic.
A dominant mobile operator and
low airtime commissions
The chances of a mobile money scheme taking root also
depend on the strength of the mobile operator within its
market, because a large market share is associated with a
larger potential customer base for cross-selling the mobile
money service, a larger network of airtime resellers that can
be converted into cash-in/cash-out agents, stronger brand
recognition and trust among potential customers, and
larger budgets to finance the heavy up-front investments
needed to deploy a new service. With a market share of
around 80 percent, Safaricom enjoyed all of these benefits
when it launched M-PESA.
Successful deployment of a mobile money application
also has a greater chance of success in countries where the
commissions that mobile operators pay airtime resellers are
relatively low. If commissions are too high, resellers will not
be attracted by the incipient cash-in/cash-out business. In
Safaricom’s case, airtime commissions total 6 percent, of
which 5 percent is passed on to the retail store. A commis-
sion of 1–2 percent on a cash-in/cash-out transaction is
plausibly attractive—the store need only believe that the
volume of the cash business will be five times the size of
the airtime business. This seems reasonable, considering
that the bulk of airtime sales are of low denominations
(around $0.25).
A reasonable base of banking infrastructure
Finally, the ability of M-PESA stores to convert cash to
e-value for customers depends on how easily they can rebal-
ance their liquidity. If bank branch penetration is too low,
rebalancing will be more difficult to achieve, because the
agent channel is forced to develop alternative cash transport
mechanisms. Thus, a mobile payment agent network
must have at its disposal at least a minimal retail banking

CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 359

 infrastructure. There appears to be a branch penetration
“sweet spot” for mobile money, where penetration is not so
high that it hampers demand for mobile money services but
not so low that agents are unable to manage their liquidity.
Because of the branch networks of Equity Bank and other
banks and microfinance institutions, Kenya is reasonably
well supplied with rural liquidity points. Even so, shortage
of cash or electronic value for M-PESA agents is a problem
in both rural and urban areas in Kenya. Mobile payment
operators in some other Sub-Saharan African countries face
more serious liquidity constraints, especially in rural areas.
Such constraints are likely to be a major factor affecting the
success of mobile services in specific country contexts.
M-PESA’S SERVICE DESIGN: GETTING
PEOPLE ONTO THE SYSTEM
While M-PESA’s rapid growth was fueled by certain country-
specific enabling conditions, the success of such an innova-
tive service hinged on the design of the service. Conducting
financial transactions through a mobile phone is not an
intuitive notion for many people, just as walking to a cor-
ner shop to make cash deposits and withdrawals may not
at first seem natural to many. To overcome this adoption
barrier, Safaricom designed M-PESA in a way that helped
people immediately grasp how they might benefit from the
service; removed barriers that might prevent people from
experimenting with the service; and fostered trust in retail
outlets that would be tasked with promoting the service,
registering customers, and facilitating cash-in/cash-out
services.
A simple message targeting a big point
of concern among the population
In very early phases of conception, Vodafone developers
thought that M-PESA would be used as a way for
customers to repay microloans. However, as Safaricom
market-tested the mobile money proposition, they shifted
the core proposition from loan repayment to helping people
make P2P transfers to their friends and family. In its com-
mercial launch, M-PESA was marketed to the public with
just three powerful words: “send money home.” In an envi-
ronment in which families were geographically split, this
message tapped into a major concern for many Kenyans—
the risks and high costs associated with sending money over
long distances. Thus, a simple marketing message turned a
basic “e-remittance” product into a must-have “killer”
application and remains the main marketing message three
360 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA
years later. Although people have proved creative in using
M-PESA for their own needs, sending money home con-
tinues to be one of the most important uses, and the num-
ber of Kenyan households receiving money transfers has
increased from 17 percent to 52 percent since M-PESA was
introduced (FSDT 2009a).
A simple user interface
The simplicity of M-PESA’s message is matched by the sim-
plicity of its user interface. The service can be launched
right from the main menu of a mobile phone, making it
easy for users to find. And because the service resides on the
phone and does not need to be downloaded from the net-
work each time it is used, the menu loads very quickly and
prompts the user for information step-by-step. For instance,
for a P2P transfer, the user is asked to enter the destination
phone number, the amount of the transfer, and the personal
identification number (PIN) of the sender. Once the infor-
mation is gathered, it is fed back to the customer for final
confirmation. Once the customer hits OK on a mobile
phone, the data is sent to the M-PESA server in a single text
message. Consolidating all information into a single mes-
sage reduces messaging costs as well as the risk that only
part of the transaction data will be sent to the server. A final
advantage is that the application can use the security keys in
the user’s SIM card to encrypt messages end-to-end, from
the user’s handset to Safaricom’s M-PESA server.
Removing adoption barriers: free to register,
free to deposit, no minimum balances
Safaricom designed M-PESA to make it as easy as possible
for customers to try the service. Customer registration,
which can be done at any M-PESA retail outlet, is quick,
simple, and free. First, the retail clerk provides a paper reg-
istration form, where the customer enters his or her name,
ID number (from Kenyan national ID, passport, military ID,
diplomatic ID, or alien ID), date of birth, occupation, and
mobile phone number. The clerk then checks the ID and
inputs the registration information into the customer’s
mobile phone. SIM cards in Kenya are now preloaded with
the M-PESA application. If the customer’s SIM card is too
old and does not contain the application, the clerk replaces
it. The customer’s phone number is not changed even if the
SIM card is.
After Safaricom receives the application, it sends both the
customer and the retail outlet an SMS confirming the trans-
action. The SMS provides the customer a four-digit start key
(one-time password), which they use to activate their
account. Customers enter the start key and their ID number,
after which they are asked to input a secret PIN of their
choice. This completes the registration process. In addition
to leading customers through this registration process, retail
clerks explain how to use the application and discuss the
costs associated with each service. This customer support
early in the process is particularly important in rural areas,
where a significant percentage of the potential user base is
unfamiliar with the functioning of mobile phones.
The minimum M-PESA deposit amount is approxi-
mately $1.25, but there is no minimum balance require-
ment. And because customers deposit money for free, there
is no immediate barrier to taking up the service. M-PESA
charges customers only for “doing something” with their
money, such as making a transfer, withdrawing money, or
buying prepaid airtime.
The ability to send money to anyone
M-PESA customers can send money to any GSM mobile
phone subscriber on the Safaricom, Zain, Orange, or YU
networks in Kenya, regardless of whether the receiving party
is an M-PESA customer. When a transfer is sent, money is
debited from the sender’s account and the recipient receives
a code by SMS, which is used to claim the monetary value at
any M-PESA store. M-PESA is thus an account-to-cash ser-
vice, with the receiver’s experience being similar to the way
Western Union works today. M-PESA’s pricing, however, is
quite different: customers pay a higher (roughly triple) P2P
charge when sending money to a non-M-PESA customer,
but at the other end the noncustomer is not charged to
receive the cash, whereas registered customers pay a cash-
out fee of at least $0.30. Safaricom developed this pricing
scheme with the understanding that the sender has power
over the recipient, and so it chose to put pressure on the
sender to require the recipient to register with M-PESA.
Furthermore, Safaricom hoped that providing noncus-
tomers with a good, no-cost first experience with M-PESA
would lure them to register for M-PESA.
Building trust in the retail network
Recognizing that M-PESA would not be rapidly adopted by
the Kenyan population unless customers had enough trust
in the M-PESA retail network that they were willing to
conduct cash-in/cash-out transactions through those out-
lets, Safaricom employed several measures to build that
trust. First, it closely linked the M-PESA brand with its own
strong corporate brand. As the mobile operator in Kenya
with the dominant share of the market (more than 80 per-
cent at M-PESA’s launch and almost as much in 2010),
Safaricom was already a broadly respected and trusted
brand, even among low-income customers. M-PESA retail
outlets are required to paint their store “Safaricom green,” a
tactic that not only gives customers confidence that the store
is acting on behalf of Safaricom but also makes it easier for
customers to locate cash-in/cash-out points.
Second, by investing heavily in store training and on-site
supervision, Safaricom ensured that customers can have a
remarkably similar experience in any retail authorized out-
let they use. This “sameness” has helped to build trust in
both the platform and the outlets and gives customers a
consistently positive view of the service. Rather than relying
on its channel intermediaries to carry out these functions in
retail shops, Safaricom chose to centralize the functions
through a single third-party vendor, Top Image. Quality is
maintained through a rating process. A Top Image repre-
sentative visits each outlet at least once per month and rates
each store on a variety of criteria, including visibility of
branding and the tariff poster, availability of cash and M-
PESA electronic value to accommodate customer transac-
tions, and quality of record-keeping.
Third, the fact that M-PESA customers receive instant
SMS confirmation of their transaction has helped them
learn by experience to trust the system. Because the receipt
confirming a money transfer includes the name and num-
ber of the recipient and the amount transferred, it allows
the sender to confirm instantly that the money was sent to
the right person—the most common source of error. In the
case of an error, the receipt can then be used to resolve the
situation.
Fourth, Safaricom requires its outlets to record all cash-
in/cash-out transactions in a paper-based, Safaricom-
created logbook. For each transaction, the store clerk enters
the M-PESA balance, the date, agent ID, transaction ID,
transaction type (customer deposit or withdrawal), value,
customer phone number, customer name, and the customer’s
national ID number. Customers are then asked to sign the
log for each transaction, which not only discourages fraud
but also gives agents a way to offer first-line customer care
for customers querying previous transactions. Each entry in
the log is written in triplicate. The top copy is kept by the
retail outlet for its own records, a second is passed on to the
store’s master agent, and the third is sent to Safaricom.
Because all information contained in the logbook (except
for the customer signature) is captured electronically by
Safaricom when the transaction is made and is available to
CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 361
 the master agents through a Web management system, the
main purpose of the agent log is not for record-keeping but
rather to provide comfort to customers who are accustomed
to having transactions recorded on paper.
Simple and transparent pricing
M-PESA pricing is transparent and predictable. There are
no customer charges for the SMSs that deliver the service.
Instead, fees are applied to the actual, customer-initiated
transactions. All customer fees are subtracted from the cus-
tomer’s account, and outlets cannot charge any direct fees.
Thus, outlets collect their commissions from Safaricom
(through their master agents) rather than from customers.
This arrangement reduces the potential for agent abuses.
Customer fees are uniform nationwide, and they are promi-
nently posted in all outlet locations in the poster shown in
figure 20.7.
M-PESA chose to specify its fees in fixed currency terms
rather than as a percentage of the transaction. This makes it
easier for customers to understand the precise cost of each
transaction and helps them think of the fee in terms of the
transaction’s absolute value (for example, sending money to
grandmother). It also helps them compare the transaction
cost against alternative and usually costlier money-transfer
arrangements (for example, the matatu fare plus travel time).

Figure 20.7 M-PESA Tariff Structure

Transaction type Transaction range (KShs) Customer charge

Minimum Maximum (KShs)

Value movement transactions

Deposit cash 100 35,000 0
Send money to a registered M-PESA user 100 35,000 30
100 2,500 75
2,501 5,000 100
Send money to a nonregistered M-PESA user 5,001 10,000 175
10,001 20,000 350
20,001 35,000 400
100 2,500 25
2,501 5,000 45
Withdraw cash by a registered M-PESA user at
an M-PESA agent outlet 5,001 10,000 75
10,001 20,000 145
20,001 35,000 170
200 2,500 30
2,501 5,000 60
Withdraw cash by registered M-PESA user at PesaPoint ATM
5,001 10,000 100
10,001 20,000 175
Withdraw cash by a non-registered M-PESA user 100 35,000 0
Buy airtime (for self or other) 20 10,000 0
Pay Bill Transactions - - 0–30
Information transactions

Show balance 1
Change secret word 0
Change PIN 20
Update menu 0
Change language 0
SIM replacement 20

Source: Safaricom.

362 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA

 M-PESA deposits are free to customers. Withdrawals of
less than $33 cost around $0.33. Withdrawal charges are
“banded” (that is, larger transactions incur a larger cost) so
as not to discourage smaller transactions. ATM withdrawals
using M-PESA are slightly more expensive than withdrawals
at a retail outlet ($0.40 versus $0.33).
P2P transfers using M-PESA cost a flat rate of around
$0.40. These transfers are where Safaricom makes the bulk
of its revenue. Thus, for a purely electronic transfer, cus-
tomers pay more than double the price they pay for the
average cash transaction ($0.17), even though the cost of
providing purely electronic transactions is lower than cash
transactions. This model reflects a notion of optimal pricing
that is based less on cost and more on customer willingness
to pay. M-PESA is still cheaper, though, than the other avail-
able mechanisms for making remote payments, such as
money transfer by the bus companies, Kenya Post’s
Postapay, or Western Union.11
Notably, M-PESA has maintained the same pricing for
transactions in its first three years of operation. This strat-
egy has helped establish customer familiarity with the ser-
vice. Safaricom has changed, however, the price of two
types of customer requests that do not involve a financial
transaction: balance inquiries (because the initial low price
generated a burdensome volume of requests) and PIN
changes (because customers are far more likely to remem-
ber their PIN if the fee to change it is higher). The volume
of both types of requests decreased substantially after
these price changes. As previously noted, the SMS confir-
mation of a transaction contains the available balance,
which also helps cut down on the number of balance
inquiries.
Liquidity of last resort at bank
branches and ATMs
From very early on, M-PESA signed up banks as agents, so
M-PESA customers could walk into any branch of those
banks to conduct cash-in/cash-out transactions. One year
after its launch, M-PESA went further and partnered with
PesaPoint, one of the largest ATM service providers in
Kenya. The PesaPoint network includes more than 110
ATMs scattered in all eight provinces of the country. Cus-
tomers can now retrieve money from any PesaPoint ATM.
To do so, they select “ATM withdrawal” from the M-PESA
menu on their mobile phone, after which they receive a one-
time ATM authorization code. They then enter that code on
the ATM keyboard to make the withdrawal. No bank card is
needed for the transaction.
M-PESA’s liquidity system is not without its challenges,
however. Because of cash flow constraints, M-PESA retail
outlets cannot always meet requests for withdrawals, espe-
cially large ones. Furthermore, the agent commission
structure discourages outlets from handling large transac-
tions. As a result, customers are sometimes forced to spread
large withdrawals over several days rather than withdraw a
lump sum, at an added cost and inconvenience. Cash flow
constraints also undermine customer trust in M-PESA as a
mechanism for high-balance, long-term saving. Use of bank
branches and ATMs to give customers a sort of liquidity
mechanism of last resort has bolstered the credibility of the
M-PESA system, however,
M-PESA’S EXECUTION STRATEGY: QUICKLY
REACHING CRITICAL MASS
Although strong services design has been a major factor in
the success of M-PESA, an appropriate execution strategy
has also been a key factor. Importantly, Safaricom recog-
nized that it would be difficult to scale M-PESA incremen-
tally, because it had to overcome three significant obstacles
common to any new electronic payment system, namely,
adverse network effects, the chicken-and-egg trap, and
trust. First, in regard to adverse network effects, the value to
the customer of a payment system depends on the number
of people connected to and actively using it. The more peo-
ple on the network, the more useful it becomes.12 While
network effects can help a scheme gain momentum once it
reaches a critical mass of customers, they also make it diffi-
cult to attract early adopters of the new technology. Sec-
ond, to grow, M-PESA had to attract customers and stores
in tandem. It is difficult, however, to attract customers
when there are few stores to serve them, and equally hard to
convince stores to sell the service when there are few
customers to be had (thus, the chicken-and-egg trap).
Thus, M-PESA needed to drive both customer and store
acquisition aggressively. Third, a company will be success-
ful in attracting customers only when prospective cus-
tomers have confidence in the reliability of the new system.
In the case of M-PESA, customers had to be comfortable
with three elements that were new at the time in Kenya: a
payment system operated by a mobile operator, using a
nonbank retail outlet to meet cash-in/cash-out needs, and
using mobile phones to access account information and
initiate transactions.
In the early stages of development of a payments system,
the three problems described above reinforce each other,
creating a significant hurdle to growth. In many cases, this
CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 363
 hurdle helps explain why many other mobile money deploy-
ments remain subscale. M-PESA, however, overcame this
hurdle through very forceful execution on two key fronts:
Safaricom made significant up-front investments in build-
ing a strong service brand for M-PESA, and it effectively
leveraged its extensive network of airtime resellers to build
a reliable, consistent retail network that served customers’
liquidity needs.
Aggressive up-front investment in
promoting the M-PESA brand
From the beginning, Safaricom sought to foster customer
trust in M-PESA and relied on existing customers to be the
prime mechanism for drawing in new customers. The task
was all the more difficult because Safaricom was not only
introducing a new product, but an entirely new product
category, to a market that had little experience with formal
financial services. Safaricom’s initial target for M-PESA
was about 1 million customers within one year of the
launch of the service, equal to 17 percent of Safaricom’s
customer base of about 6 million customers at the time,
according to Safaricom company results for the year end-
ing March 2007.
National launch at scale. After small pilots involving
less than 500 customers,13 M-PESA launched nationwide,
increasing the likelihood that the service could reach a
critical mass of customers in a short time frame. At launch,
Safaricom had 750 stores that covered all of Kenya’s 69 dis-
trict headquarters. The launch was a massive logistical
challenge that led to a great deal of customer and store con-
fusion and, during the first few months, delays of several
days in reaching customer service hotlines. User and store
errors were frequent since everyone was new to the service.
But the gamble paid off. Logistical problems subsided after
a few months, leaving strong brand recognition and top-
of-mind awareness among large segments of the popula-
tion. The service outran first-year growth targets, quickly
turning network effects in its favor as new customers begat
more customers and turned M-PESA into a compelling
business proposition for more stores.
An appropriate marketing mix. Initial M-PESA market-
ing featured and targeted urban, relatively wealthy city
dwellers with a need to “send money home.” The choice
of this demographic as the initial customer created an aspi-
rational image for prospective customers and avoided giv-
ing the impression that M-PESA was a low-value product
364 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA
aimed at the poor. Over time, the marketing moved from
young, up-market urban dwellers with desk jobs to more
ordinary Kenyans with lower-paid professions.
M-PESA’s launch was associated with significant up-
front investment in television and radio marketing,14 but
there was also intense outreach through road shows, in
which company agents traveled around the country signing
people up, explaining the product, and demonstrating how
to use it. Over time, as people became more familiar with
M-PESA, it was no longer necessary to do this kind of
hands-on outreach. Television and radio marketing was
largely replaced by the omnipresent M-PESA branding at
all outlets, supported by a few large billboards. Newer ads
feature a general emotional appeal, with a wider range of
services indicated.
A scalable distribution channel
From the time Safaricom launched M-PESA, it under-
stood that the primary role of the mobile phone is to
enable the creation of a retail outlet–based channel for
cash-to-digital value conversion. It also understood that
for this cash-to-digital conversion to be broadly available
to the bulk of the population, Safaricom had to develop a
channel structure that could support thousands of
M-PESA stores spread across a broad geographical area.
To achieve this, Safaricom built four elements into its
channel management execution strategy: engaging inter-
mediaries to help manage the individual stores, thereby
reducing the number of direct contacts it had to deal
with, ensuring that outlets received sufficient incentives
to actively promote the service, maintaining tight control
over the customer experience, and developing several dif-
ferent methods for stores to rebalance their stocks of cash
and e-value.
Two-tier channel management structure. Safaricom
created a two-tier structure with individual stores (sub-
agents, in Safaricom’s parlance) that depends on master
agents (referred to by Safaricom as agent head offices).
Agent head offices maintain all contact with Safaricom, and
perform two key functions: liquidity management (buying
and selling M-PESA balance from Safaricom and making it
available to individual stores under their responsibility),
and distribution of agent commissions (collecting the com-
mission from Safaricom based on the overall performance
of the stores under them and remunerating each store).
Individual stores are either directly owned by an agent head
office or working for one under contract.
Giving stores incentives. Retail outlets will not maintain
sufficient stocks of cash and electronic money unless they
are adequately compensated for doing so. Hence, Safaricom
pays commissions to agent head offices for each cash-
in/cash-out transaction conducted by stores under their
responsibility. Although Safaricom does not prescribe the
commission split between agent head offices and stores,
most agent head offices pass on 70 percent of commissions
to the store.15 For deposits under $33, Safaricom pays a
$0.133 commission, of which $0.074 goes to the store after
tax. For withdrawals, Safaricom pays $0.200, of which
$0.111 goes to the store. So, assuming equal volumes of
deposits and withdrawals, the store earns $0.092 per trans-
action. Assuming the store conducts 60 transactions per day,
it earns around $5.50—almost twice the prevailing daily
wage for a clerk in Kenya.
Recall that Safaricom charges customers $0.333 on a
round-trip savings transaction (free deposit plus $0.33 for
the withdrawal), which is in fact equal to what the channel
receives ($0.13 on the deposit plus $0.20 for the withdrawal).
So, assuming equal volumes of deposits and withdrawals,
Safaricom does not make money on cash transactions. It
merely “advances” commissions to the channel when cus-
tomers make deposits and recoups them when customers
withdraw. As noted, Safaricom generates the bulk of its rev-
enue from services for which customer willingness to pay is
the highest—electronic, P2P transactions.
Because store revenues are dependent on the number of
transactions they facilitate, Safaricom has been careful not
to flood the market with too many outlets, lest it depress the
number of customers per agent. Instead, it has maintained
balanced growth in the number of outlets relative to the
number of active customers, resulting in an incentivized
and committed agent base.
Maintaining tight control over the customer
experience. Safaricom also recognized early on that cus-
tomers need to have a good experience at the retail stores
offering M-PESA services, where the bulk of transactions
take place. To ensure that it maintained control over the
customer experience, it has not relied on agent head offices
to perform all channel management functions. Instead, it
concentrated the evaluation, training, and on-site supervi-
sion of stores in a single outsourcing partner, Top Image.
The more routine and non-customer-facing store support
activities, such as liquidity management and distribution of
store commissions, are left to a large pool of agent head
offices. Through its contract with Top Image, however,
Safaricom retained direct, centralized control over key
elements of the customer experience: store selection, train-
ing, and supervision.
Developing multiple store liquidity management
methods. By far the biggest challenge faced by M-PESA
stores is maintaining enough liquidity, in terms of both
cash and e-float,16 to be able to meet customer requests. If
they take too many cash deposits, stores will find them-
selves running out of e-float with which to facilitate fur-
ther deposits. If they do too many withdrawals, on the
other hand, they will accumulate e-float but will run out of
cash. Thus, stores frequently undertake liquidity manage-
ment efforts.
The M-PESA channel management structure was con-
ceived to offer stores three methods for managing liquid-
ity. Two of these place the agent head office in a central
role, with the expectation that it will “recycle” e-float
between locations experiencing net cash withdrawals (that
is, accumulating e-float) and locations with net cash
deposits (that is, accumulating cash). In the first method,
the agent head office provides direct cash support to stores.
The store clerk comes to the agent head office to deliver or
offload cash, or the agent head office sends cash runners to
the store to perform these functions.
In the second method, the agent head office and stores
use their respective bank accounts. If the store has excess
cash and wants to buy M-PESA e-float from the agent head
office, the store deposits the cash into the account of the
agent head office at the nearest bank branch or ATM. Once
the agent head office confirms receipt of the funds into its
account, it transfers M-PESA e-float to the store’s M-PESA
account. If the store wants to sell e-float to obtain cash, the
store transfers M-PESA e-float to the agent head office. The
agent head office then deposits (or transfers) money into
the store’s bank account, after which the store can withdraw
the cash at the nearest branch or ATM.
In the third method, stores interact directly with a bank
that has registered as an M-PESA “superagent.” Under this
method, the agent head office does not get involved in liq-
uidity management. Instead, stores open an account with a
participating superagent bank. To rebalance their cash,
stores deposit and withdraw cash against their bank account
at the nearest branch or ATM of the bank. The store then
electronically buys and sells e-float in real time against its
bank account. From a store’s perspective, one drawback of
using a bank-based superagent mechanism is that it can
only be used during banking business hours.
In all cases, the e-float–cash nexus will remain the key
constraint to further development of M-PESA, because it
CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 365
 requires the physical movement of cash around the country
and is thus the least scalable part of the system.
M-PESA’S FUTURE EVOLUTION
The experience of M-PESA demonstrates how powerful a
payment network that offers convenience at an affordable
cost can be once a critical mass of customers is reached. It also
shows that achieving critical mass requires a service design
that removes as many adoption barriers as possible, together
with significant investment in marketing, branding, and
agent network management. The Kenyan experience also sug-
gests that several country-level factors must be aligned to set
the scene for successful mobile money development,
including the labor market profile (demand for remittances
generated by rural-urban migration), the quality of available
financial services, support from the banking regulator, and
the structure of the mobile communications market (domi-
nant mobile operator and low airtime commissions).
While M-PESA has been more successful than anyone
could have imagined at its launch, the model still has sub-
stantial room for development. A threefold wish list for
M-PESA could be delineated as follows: further main-
streaming of M-PESA’s regulatory treatment, pricing that
opens up a much larger market of microtransactions, and
expanding M-PESA so that customers have access to a
broader range of financial services.
Mainstreaming M-PESA’s regulatory treatment
M-PESA’s regulatory treatment as a payments vehicle needs
to be formalized so that it can become regulated in the most
appropriate way. To this end, the Central Bank of Kenya is
backing a new payments law that would cover M-PESA trans-
actions (as of this writing, the draft had not yet been
approved by the Kenyan parliament). The Central Bank of
Kenya is also in the process of finalizing agent banking regu-
lations that would allow commercial banks to use retail out-
lets as a delivery channel for financial services. Banks are quite
reasonably complaining that they could not replicate the
M-PESA service themselves because they are not currently
allowed to undertake customer transactions through agent
networks on their own. Allowing both banks and M-PESA to
operate such agent networks would level the playing field.
Pricing that enables smaller payments
M-PESA’s current pricing model is not conducive to small
transactions. The fee for a $10 P2P transfer plus withdrawal,
366 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA
for example, is approximately 7 percent of the amount of
the transaction ($0.40 for the transfer plus $0.33 for the
withdrawal). Adjusting M-PESA’s current pricing model to
account for smaller-denomination transactions would have
two advantages. First, it would make the service more acces-
sible to the poor, for whom pricing is now too high given
their transactional needs. Such a reduction would allow
Safaricom to maintain customer growth once saturation
starts to set in at current pricing. Second, a pricing adjust-
ment would allow customers to use M-PESA for their daily
transaction needs, and in particular to save on a daily basis,
which would be beneficial to those who are paid daily.
A reduction in customer prices could come about in sev-
eral ways. There is room for “tranching” the P2P fee of
$0.40, for example, so that the price of smaller, or more fre-
quent, transactions becomes more affordable. For cash
transactions, one way to enable lower fees would be to
establish street-level M-PESA subagents who would offer
lower costs and commissions than store-based agents. Sub-
agents would be a kind of “e-susu collector,” operating with
small working capital in order to aggregate small customer
transactions. Subagents would use normal M-PESA retail
outlets to rebalance their cash and M-PESA stored value.
The key principle here is that segmentation of customers
would go hand-in-hand with segmentation of agents.
Linking with banks and other institutions
While some customers use M-PESA as a saving device, the
service still falls short of being a useful method of saving
for most poor people. The fact that the average balance of
M-PESA accounts was less than $3 in early 2009 is partly a
“large number” problem: if 900,000 people used M-PESA
to save, that means 10 percent of users use the service to
save, and that the average savings balance is diluted because
it takes into account all M-PESA users rather than only
users who save. But the fundamental problem is that there
is still a lot of conversion of electronic value back into cash.
This can be attributed to a combination of factors:
■ Lack of marketing. Safaricom does not want to publicly
promote using M-PESA as a saving tool for fear of pro-
voking the Central Bank Kenya to regulate it more
tightly.
■ Customer pricing. The flat fee of around $0.33 for with-
drawals under $33 means that small withdrawals carry a
large relative fee.
■ Product design. M-PESA works very much like an elec-
tronic checking account and does not offer structured
 saving products that may help people build discipline
around savings.
■ Inflation. M-PESA does not pay interest. In an environ-
ment with 15 percent inflation (during its first full year
of operation in 2008), saving may be too onerous for
much of the Kenyan population.
■ Trust. M-PESA deposits are not supervised by the Cen-
tral Bank of Kenya. And unlike payments, where trust
can be validated experientially in real time, saving
requires garnering the trust of customers over a longer
period of time.
■ Privacy. People may want more privacy for their saving
behavior than an agent provides.
■ Excess liquidity. The 16,000 M-PESA cash-in points in
Kenya are also 16,000 cash-out points. The ubiquity of
M-PESA agents may make it too easy for customers to
cash out their funds, thus limiting their ability to accu-
mulate large balances.
Rather than expecting Safaricom to develop and market
more comprehensive savings services, M-PESA should sup-
port saving by linking to banks. M-PESA could then become
a massive transaction acquisition network for banks rather
than an alternative to them. That said, Safaricom is begin-
ning to connect with banks. In September 2009, for exam-
ple, Family Bank and M-PESA established a connection
allowing customers to transfer money from M-PESA to
their Family Bank account using M-PESA’s bill pay func-
tion. This connection follows a successful pilot of loan
repayments via M-PESA’s bill pay function.
M-PESA would also benefit from establishing further
links with institutions beyond banks, such as billers, dis-
tributors, and employers. By promoting M-PESA as a
mechanism for distributing salaries and social welfare pay-
ments, enabling payments across supply chains, and paying
bills, the need for cash-in and cash-out services would be
minimized, and, as a result, a key component of transaction
costs could be reduced. Savings balances may also be higher
if people received payments directly into their accounts
rather than in cash, and if they had other useful things to
do with their money in electronic form.
CONCLUDING THOUGHTS: HOW M-PESA
CAN REINVIGORATE FINANCIAL
INCLUSION EFFORTS
Imagine a world where there are no banks where you live.
The nearest branch is 10 kilometers away, and it takes you
almost an hour to get there by foot and bus. With waiting
times at the branch, a trip to the bank and back may take
two hours—approximately a quarter of your working day. A
bus fare of only $0.50 to get to the bank may well represent
one-quarter of your income on a good day. With the bank
fees included, each banking transaction costs you the equiv-
alent of almost half a day’s wages. It would be like charging
someone with an average income in the United States
something like $50 for each ATM transaction. Then, imag-
ine a world without credit instruments or electronic pay-
ments. No checks, no credit cards, no money orders, no
direct debits, no Internet banking. All your transactions are
done in cash or, worse, by bartering goods. All exchanges
are physical, person-to-person, hand-to-hand. Consider
the hassle and the risk of sending money to distant rela-
tives, business partners, or banks.
How would you operate in such a world? A recent book,
Portfolios of the Poor, documents how poor people cope
(Collins et al. 2009). Some people save to “push” excess
money from today to tomorrow, some people borrow to
“pull” tomorrow’s money to fund necessary expenses today.
They store cash in the home to meet daily needs, they leave
it with a trusted friend for emergencies, they buy jewelry
because that represents a future for their children, they pile
up bricks for the day when they can build an extra room in
their house. They make contributions to a savings group
with a circle of friends to build up a pot of money, and one
day it will be their turn to take that pot home to buy new
clothes. They borrow from friends, seek advances from their
employers, pawn their jewelry, and borrow from a high-
interest moneylender.
Lack of good financial options is undoubtedly one of the
reasons why poor people are trapped in poverty. In many
cases, poor people cannot sustain themselves or aspire to
earn higher incomes because they are not able to invest in
better farming tools and seeds to enhance their productiv-
ity, start a microenterprise, or even take the time to search
for better-paying employment opportunities. Their income
is volatile, often fluctuating daily, and without reliable ways
of pushing and pulling money between good days and bad
days, they face stark choices such as pulling their children
out of school or putting less food on the table during bad
patches. Without good financial tools, they also may not be
able to cope with shocks that periodically set them back.
Most of these shocks are foreseeable, if not entirely pre-
dictable: a drought, ill health, and lifecycle events such as
marriage and death.
Cash is the main barrier to financial inclusion. As long as
poor people are able to exchange value only in cash—or
worse, physical goods—they will remain too costly for
CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 367
 formal financial institutions to address in significant
numbers. Few banks are willing to build the costly infra-
structure necessary for collecting low-value cash deposits
and redeeming savings back into small sums of cash in
low-income or rural areas. But once poor people have access
to cost-effective electronic means of payments such as
M-PESA, they could, in principle, become profitable to
financial institutions. Although M-PESA itself does not
constitute financial inclusion, it does provide a glimpse of a
commercially sound, affordable, and effective way to offer
financial services to all.
NOTES
1. See Hughes and Lonie (2009) for a historical account of
the M-PESA service, Mas and Morawczynski (2009) for a
fuller description of the service, and Mas and Ng’weno
(2009) for the latest accomplishments of M-PESA.
2. A SIM card is a smart card found inside mobile phones
that are based on the Global System for Mobile communi-
cations (GSM) family of protocols. The SIM card contains
encryption keys, secures the user’s persona identification
number on entry, and drives the phone’s menu. SMS is a
data messaging channel available on GSM phones.
3. These amounts use an exchange rate of $1 to 75 Kenyan
shillings.
4. Kenya has a population of nearly 40 million. GDP per
capita is $1,600, 78 percent of people live in rural areas, and
19 percent of adults have access to a formal bank account.
See FSDT (Financial Sector Deepening Trust) (2009a) for
financial access data derived from the FinAccess survey, a
nationally representative survey of 6,600 households con-
ducted in early 2009.
5. The 2009 FinAccess survey (FSDT 2009a, p. 16) con-
firms that 40 percent of adults have used M-PESA.
6. See Safaricom (2009) for key monthly M-PESA statis-
tics. Additional figures are taken from Safaricom’s published
half-year results for the period ending September 2009 and
Central Bank of Kenya reports.
7. Although the number of P2P transactions per customer
has been rising steadily, it remains quite low, probably still
less than two transactions per month.
8. For more detailed accounts of the M-PESA story, see
Heyer and Mas (2009) on the country factors that led to
M-PESA’s success, Mas and Morawczynski (2009) on
M-PESA’s service features, Mas and Ng’weno (2010) on
Safaricom’s execution, and Mas (2009) on the economics
underpinning branchless banking systems.
9. For a fuller analysis of the use of mobile money for
domestic remittances in Kenya, see Ratan (2008) and
Morawczynski (2008).
368 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA
10. Although the Central Bank of Kenya Act was amended
in 2003 to give the central bank broad oversight of pay-
ment systems, the operational modalities for this oversight
have not been implemented; they are pending approval of
a new National Payments System Bill that has languished
in parliament.
11. Morawczynski and Pickens (2009) find that sending
K Sh 1,000 through M-PESA is 27 percent cheaper than the
post office’s PostaPay and 68 percent cheaper than sending
it by a bus company.
12. Network effects are commonly illustrated with reference
to fax machines: the first set of people who bought a fax
machine did not find them very useful, because they were
not able to send faxes to many people. But as more people
bought fax machines, everyone’s faxes became increasingly
useful. Network effects are sometimes called demand-side
economies of scale to emphasize that scale affects the value
of the service to each customer. This distinguishes it from
supply-side economies of scale, which refer to situations in
which the average cost per customer fall as volume increases.
Davidson (2009) discusses implications of network effects
for mobile money.
13. Safaricom’s earliest pilot project conducted in 2004–05
revolved around microloan repayments and involved the
Commercial Bank of Africa, Vodafone, Faulu Kenya, and
MicroSave.
14. In a survey of 1,210 users in late 2008, 70 percent of
respondents claimed they first heard about M-PESA from
television or radio advertisements (FSDT 2009b).
15. Safaricom would like the split to be 20 percent/80 per-
cent, meaning that more of the commission is passed on to
the retail outlet.
16. E-float is the balance of money that an agent has in his
M-PESA account, which he can electronically transfer to
customers in exchange for cash.
REFERENCES
Camner, Gunnar, and Emil Sjöblom. 2009. “Can the Success of
M-PESA be repeated? A Review of Implementations in
Kenya and Tanzania.” Valuable Bits. http://mobileactive.org/
files/file_uploads/camner_sjoblom_differences_ke_tz.pdf.
Collins, Daryl, Jonathan Morduch, Stuart Rutherford, and
Orlanda Ruthven. 2009. Portfolios of the Poor: How the
World’s Poor Live on $2 a Day. Princeton, NJ: Princeton
University Press.
Davidson, Neil. 2009. “Tactics for Tipping Markets: Influ-
ence Perceptions and Expectations.” Mobile Money for
the Unbanked Blog, GSM Association, November 15.
Financial Sector Deepening Trust (FSDT). 2007a. “Financial
Access in Kenya: Results of the 2006 National Survey.”
FSDT, Nairobi.
———. 2007b. “Key Findings of the FinScope Survey in
Tanzania in 2006.” FSDT, Nairobi.
———. 2009a. “FinAccess National Survey 2009: Dynamics
of Kenya’s Changing Financial Landscape.” FSDT, Nairobi.
———. 2009b. “Research on Mobile Payments Experi-
ence: M-PESA in Kenya.” Unpublished paper, FSDT,
Nairobi.
GSM Association. 2009. “Wireless Intelligence Database.”
http://www.wirelessintelligence.com.
Heyer, Amrik, and Ignacio Mas. 2009. “Seeking Fertile
Grounds for Mobile Money.” Unpublished paper.
Hughes, Nick, and Susie Lonie. 2009. “M-PESA: Mobile
Money for the Unbanked.” Innovations (special edition
for the Mobile World Congress 2009).
Isaacs, Leon. 2008. IAMTN presentation. MMTA Confer-
ence, Johannesburg, May.
Jack, William, and Tavneet Suri. 2009. “Mobile Money: the
Economics of M-PESA.” Unpublished paper.
Juma, Victor. 2009. “Family Bank Offers New Service Link-
ing Accounts to M-PESA.” Business Daily, December 18.
Kimenyi, Mwangi, and Njuguna Ndung’u. 2009. “Expand-
ing the Financial Services Frontier: Lessons from Mobile
Phone Banking in Kenya.” Washington, DC: Brookings
Institution.
Kinyanjui, Kui. 2009. “Yu Launches New Mobile Cash
Transfer Platform.” Business Daily, 16 December.
Mas, Ignacio. 2008a. “M-PESA vs. G-Cash: Accounting for
their Relative Success, and Key Lessons for other Coun-
tries.” Unpublished paper, CGAP, Washington, DC.
———. 2009. “Exploring the Usage and Impact of Trans-
formational M-banking: The case of M-PESA in Kenya.”
Unpublished paper.
Morawczynski, Olga, and Gianluca Miscione. 2008. “Exam-
ining Trust in Mobile Banking Transactions: The Case of
M-PESA in Kenya.” In Social Dimensions of Information
and Communication Technology Policy: Proceedings of the
Eighth International Conference on Human Choice and
Computers (HCC8), IFIP TC 9, Pretoria, South Africa,
September 25–26, 2008. ed. C. Avgerou, M. Smith, and
P. van den Besselaar, 287–298 (Boston, MA: Springer).
Morawczynski, Olga, and Mark Pickens. 2009. “Poor People
Using Mobile Financial Services: Observations on
Customer Usage and Impact from M-PESA.” CGAP
Brief, CGAP, Washington, DC. http://www.cgap.org/p/
site/c/template.rc/1.9.36723/.
Okoth, Jackson. 2009. “Regulator Gives M-PESA a Clean Bill
of Health.” The Standard, 27 January.
Okuttah, Mark. 2009. “Safaricom Changes Method of Recruit-
ing M-PESA Agents.” Business Daily, December 223.
Ratan, A. L. 2008. “Using Technology to Deliver Financial
Services to Low-Income Households: A Preliminary
Study of Equity Bank and M-PESA Customers in Kenya.”
Microsoft Research Technical Report.
Safaricom. 2009. “M-PESA Key Performance Statistics.”
http://www.safaricom.co.ke/fileadmin/template/main/
images/MiscUploads/M-PESA%20Statistics.pdf.
Suri, Tavneet, and William Jack. 2008. “The Performance
and Impact of M-PESA: Preliminary Evidence from a
Household Survey.” Unpublished paper.
CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 369
 CHAPTER 21
Independent Power Projects in
Sub-Saharan Africa: Determinants
of Success
Anton Eberhard and Katharine Nawal Gratwick
t the beginning of the 1990s, virtually all major
A
power generation providers in Sub-Saharan
Africa were financed by public coffers, including
concessionary loans from development finance institutions
(DFIs). These publicly financed power generation assets
were considered one of the core elements of state-owned,
vertically integrated power systems. A confluence of fac-
tors, however, brought about a significant change in the
ensuing years.
With insufficient public funds for new power generation
and decades of poor performance by state-run utilities,
African countries began to adopt a new model for their
power systems, influenced by pioneering reformers in
Chile, Norway, the United Kingdom, and the United
States.1 Urged on by multilateral and bilateral develop-
ment institutions, which were withdrawing funding from
state-owned projects, a number of countries adopted
plans to unbundle their power systems and introduce pri-
vate participation and competition. Independent power
projects (IPPs)—privately financed, greenfield generation
supported by nonrecourse or limited-recourse loans and
with long-term power purchase agreements (PPAs) with
the state utility or another off-taker—thus became a pri-
ority in overall power sector reform (World Bank 1993;
World Bank and USAID 1994). IPPs were considered a
solution to persistent supply constraints, and they also
Anton Eberhard is a professor and Katharine Nawal Gratwick is a PhD graduate at the University of Cape Town Graduate School of Business.
had the potential to benchmark state-owned supply to
gradually introduce competition (APEC Energy Working
Group 1997).
In 1994 Côte d’Ivoire became one of the first African
countries to attract a foreign-led IPP to sell power to the
national grid under long-term contracts with the state util-
ity. Ghana, Kenya, Nigeria, Senegal, Tanzania, and Uganda,
among others, opened their doors to foreign and local
investors in their power sectors shortly thereafter.
Although IPPs were considered part of larger power sec-
tor reform programs in Sub-Saharan Africa, the reforms
were not far-reaching. In most cases, state utilities remained
vertically integrated and maintained a dominant share of
the power generation market, while private investors were
allowed to operate only on the margin of the sector.2 Policy
frameworks and regulatory regimes, necessary to maintain a
competitive environment, were limited. International com-
petitive bids (ICBs) for IPPs were often not conducted
because of tight time frames, resulting in limited competi-
tion for the market and, because of the long-term PPAs, no
competition in the market. These long-term PPAs, along
with government guarantees and security arrangements
such as escrows and liquidity facilities, exposed countries to
significant exchange rate risks. Finally, while Africa has seen
continued private participation in greenfield electricity proj-
ects, that progress has been erratic, with 2007 representing
371
 the zenith, largely because of the financial close of one large
project, Bujagali.
Several factors explain the recent trends in investment
in Africa’s power sector. First, private sector firms were
deeply affected by the Asian and subsequent Latin Ameri-
can financial crises in the late 1990s and early 2000s. The
Enron collapse and its aftershocks also influenced U.S.-
and European-based firms to reduce risk exposure in
developing-country markets and to refocus on core activi-
ties at home. The financial crisis of 2008 and 2009 has also
had a toll. Furthermore, DFIs began to reconsider their
position of restricted infrastructure investment, a model
that was predominant throughout the 1990s.3 As conces-
sionary funding became available again, many countries
opted for a hybrid solution—part public, part private.
Kenya represents one of the clearest examples of such a
hybrid, with KenGen, the state-owned generator, building
alongside IPPs, with support from DFIs.
Despite this revival of concessionary lending, power
sector investments have been insufficient in addressing
Sub-Saharan Africa’s power needs. Only 25 percent of the
population currently has access to electricity, and poor
supply is the rule, not the exception. The cost of meeting
Africa’s power sector needs is estimated at $40.8 billion per
year, equivalent to 6.35 percent of Africa’s 2005 gross
domestic product (GDP). Approximately two-thirds of the
spending needed is for capital investment ($26.7 billion
per year), and the remainder for operations and mainte-
nance (O&M). Current spending on power infrastructure
totals approximately $11.6 billion per year. Approximately
80 percent of existing spending is domestically sourced
from taxes or user charges. The remainder is split among
official development assistance (ODA) financing (6 percent
of the total); other sources, mainly China (9 percent); and
private sector investment (4 percent). Tackling existing util-
ity inefficiencies, including system losses, underpricing,
undercollection of revenue, and overstaffing, would make
an additional $8.24 billion available, but a funding gap of
about $21 billion would still remain (Eberhard and
Shkaratan 2010).
Closing Africa’s power infrastructure funding gap
inevitably requires undertaking reforms to reduce or elimi-
nate system inefficiencies. This will help existing resources
go farther and create a more attractive investment climate
for external and private finance, which still has growth
potential. With the original drivers for market reform still
present, future private sector involvement appears inevitable.
Approximately 20 grid-connected IPPs, each in excess of
40 megawatts (MW) and with long-term PPAs with the
372 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
largely state-run utilities, have been developed in Sub-
Saharan Africa as of early 2010 (table 21.1). In total, about
four gigawatts (GW) of IPP capacity has been added. With
few exceptions,4 these IPPs represent a small fraction of
total power generation capacity and have mostly comple-
mented incumbent state-owned utilities.
Nevertheless, IPPs are an important source of new
investment in the power sector in a number of African
countries. In Togo, for example, Centrale Thermique de
Lomé, the country’s first IPP, will triple the country’s
installed capacity.
The majority of IPP contracts in Sub-Saharan Africa
have been upheld (namely, CIPREL and Azito in Côte
d’Ivoire, Takoradi II in Ghana, Iberafrica, Tsavo, OrPower 4,
and Rabai in Kenya, Afam VI and Aba Integrated in Nigeria,
and Namanve in Uganda). Although the contracts of the
two Senegalese projects, GTi Dakar and Kounoune I, remain
largely intact, there are reports of changes in fuel supply
arrangements. A number of other IPPs, such as Bui Hydro
in Ghana, Bujagali in Uganda, and Centrale Thermique de
Lomé in Togo, have reached financial closure and are under
construction.5 Kenya is in the process of negotiating three
more IPPs after an international competitive tender.
For all of this progress, however, there have been sev-
eral high-profile IPP mishaps in Sub-Saharan Africa. Two
projects—AES Barge in Nigeria and Independent Power
Tanzania Limited (IPTL) in Tanzania—are in arbitration.
The costs of Songas, in Tanzania, meanwhile, have
escalated as a result of the unplanned, and later disputed,
contracting of IPTL; Songas’s capacity charges were later
reduced after the government agreed to buy down the
accumulated allowance for funds used during construction
costs. A dispute about escalating investment costs also
marked the Okpai project in Nigeria. In Kenya changes
may be made in the contracts of OrPower 4, which
reduced its tariff for the second phase of the plant.
Another Kenyan project, Westmont, had an initial seven-
year contract that was not renewed. The other early IPP in
Kenya, IberAfrica, had its contract renewed but with
much lower capacity charges.
Following contract changes, IPPs have generally gone on
to make a significant contribution to the countries’ power
generation—the main exceptions being Westmont, which
ceased operation, and IPTL, which has operated intermit-
tently during its arbitration proceedings. Another high-
profile failure was the nontransparent procurement process
surrrounding the Richmond/Dowans plant in Tanzania,
which has not been allowed to operate since corruption
charges were filed. Furthermore, there is evidence of stalling
 Table 21.1 General Project Specifications of Sub-Saharan African IPP Projects
Country/ Length of Commercial
project Size (MW) Fuel/cycle Contract type contract (years) Project tender operating data
East Africa
Kenya
Westmont 46 Kerosene/gas condensate/gas Turbine (barge-mounted) BOO 7 1996 1997
Iberafrica 109a HFO/medium-speed diesel engine BOO 7/15/25 1996,1999, 2008 1997, 2000, 2009
OrPower 4 48 Geothermal BOO 20 1996 2000, 2009
Tsavo 75 HFO/medium-speed diesel engine BOO 20 1995 2001
Rabai 90 HFO BOOT 20 2006 2009
Tanzania
IPTL 100 HFO/medium-speed diesel engine BOO 20 1997 1998
Songas 180 Natgas/open cycle BOO 20 1994 2004
Uganda
Namanve 50 HFO BOOT 6 — 2008
Bujagali 250 Hydro BOT 30 2005b —
West Africa
Côte d’Ivoire
CIPREL 210 Natgas/open cycle BOOT 19 1993 1995
Azito 288c Natgas/open cycle BOOT 24 1996 2000
Ghana
Takoradi II 220d Light crude/single cycle BOOT 25 1998 2000
Sunon Asogli 200 Combustion engine BOO 20 2007 —
Bui Hydroe 400 Hydro BOO — 2005 —
Nigeria
AES Barge 270 Natgas/open cycle (barge-mounted) BOO 20 (2 parts) 1999 2001
Okpai 450 Natgas/combined cycle BOO 20 2001 2005
Afam VI 630 Natgas/combined cycle BOO 20 2000 2007
Aba Integratedf 140 Natgas/open-cycle BOO 20/15 2005 —
Senegal
GTi Dakar 52 Diesel/Nafta BOOT 15 1996 1999
Kounoune I 68 HFO BOO 15 2003 2008
Togo
Centrale Thermique de Lomé 100 Triple fuel (natgas/HFO/diesel) BOOT 25 — 2010

Source: Authors.
Note: Projects included here are greater than 40 MW; have reached financial close; and are under construction, operational, or concluded. BOO = build-own-operate contract; BOOT = build-own-
operate-transfer contract; HFO = Heavy fuel oil. — = Not available.
a. Iberafrica has been developed in three stages, with 44, 12 and 52.3 MW, totaling 109 MW at time of writing.
b. The first phase of Bujagali’s conceptualization, spanning the mid-1990s until 2003 and involving AES, is not covered in this article. It should be noted that the project did not reach financial close
during this time. Authors report only on the project from its second phase, starting in 2005.
c. The initial project concept included specifications to raise capacity to 420 MW.
d. Although the initial project concept included specifications to add a second phase of 110 MW and convert to combined cycle, lack of funding has limited the completion of this phase.
e. The Bui project was initiated in the 1960s but aborted after the coup in Ghana in 1966. The project was reconsidered several times in the decades that followed. In 2005 the government of Ghana
signed a memorandum of understanding with the Chinese firm Sino Hydro, and the plant is expected to be online in 2012.
f. Aba Integrated is privately financed but is not a classic IPP in that ownership will extend to the off-taker.
373
 in Takoradi II’s second phase and in Sunon Asogli, which as
of early 2010 had no gas supply—although efforts to rectify
this situation are under way.
COUNTRY-LEVEL FACTORS AFFECTING
THE IPP MARKET
Several elements have contributed to the success of IPP
projects in Sub-Saharan Africa: a favorable investment cli-
mate; new policy frameworks and regulation; the linking of
planning, procurement, and contracting; and low-cost fuel
and secure fuel contracts.
Favorable investment climate
Even though a investment climate is not the only factor in
influencing IPP outcomes, it sets the stage for negotia-
tions and contract terms and helps explain the initial
imbalance in some of the Sub-Saharan African cases. Host
countries with a strong investment profile attracted more
investors and ultimately were able to negotiate deals with
more favorable terms than countries with weak invest-
ment conditions. Because all countries and entities com-
pete for capital, a risk-reward balance needs to be offered
that will attract investors and lenders. That “balance”
starts with a stable and predictable investment environ-
ment (Rudo 2010a).
Of the Sub-Saharan countries whose IPPs are examined
in this chapter, none has an investment-grade sovereign
credit rating. Of the five countries that have received a spec-
ulative rating (Ghana, Kenya, Nigeria, Senegal, and Uganda),
all except Ghana received their rating after the first IPP
deals were signed. Kenya’s investment climate was defined, at
the time, by its aid embargo in the mid-1990s. Tanzania is also
worth mentioning in this context. Throughout the 1990s, all
export credit agencies were off-cover in Tanzania. Foreign
commercial banks were not willing to lend, because there
was no track record of successful repayment of commercial
loans. Consequently, traditional, project-financed IPP deals
in this climate were limited. In contrast in North Africa, the
Arab Republic of Egypt, Morocco, and Tunisia have all
attained either a credit rating of investment grade or one
notch below and have experienced successful IPPs
(Gratwick and Eberhard 2008a). While credit enhance-
ments and security arrangements have differed broadly
between North and Sub-Saharan Africa, interestingly,
incentives offered to investors in IPPs were relatively similar
in both regions, though there was some variety with regard
to tax breaks. For instance, nearly all projects appear to have
374 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
benefited from both customs and value added tax (VAT)
exemptions during construction, as well as the right to full
repatriation of profits. Currency conversion was also pro-
vided for virtually all projects. In East Africa, Tanzania pro-
vided a tax holiday of five years, but Kenya extended this
benefit only until plants were commissioned. Although
one would expect investment incentives to increase with
perceived risk, such a pattern is not apparent.
New policy frameworks and regulation
While all eight Sub-Saharan African countries examined in
this chapter have introduced legislation allowing for pri-
vate sector power generation, few have actually formulated
and implemented a clear and coherent policy framework
for procuring IPPs. Thus, there is abundant evidence of
tentative experimentation with private power that does
not always lead to a sustained opening of the market for
private investment. Furthermore, long-term power pur-
chase agreements have the potential to constrain future
wholesale competition, although means to transition to
wholesale competition with IPPs have been identified
(Woolf and Halpern 2001). In addition, state-owned utili-
ties are rarely exposed to market costs of capital, and direct
comparisons of state-owned utilities’ costs and IPPs’ costs
are often difficult to discern.
The standard reform model for power sector reform—
namely, unbundling of generation, transmission, and dis-
tribution; as well as the introduction of competition and
private-sector participation at both the generation and
distribution level—is not being fully adopted anywhere in
Sub-Saharan Africa (UNEP and UNECA 2006; Malgras,
Gratwick, and Eberhard 2007a; Gratwick and Eberhard
2008b). Most incumbent national utilities are state-owned6
and in a dominant position. However, elements of the
reform model have been adopted: for example, Kenya has
unbundled power generation from its national transmission
and distribution utility; Uganda and Ghana have unbun-
dled generation, transmission, and distribution; and
Uganda and Côte d’Ivoire have introduced private conces-
sions. The private sector has also invested in IPPs in many
countries. In general, there has been competition for the
market, but not ongoing competition in the market in
terms of customer choice. In effect, hybrid power markets
have emerged across Africa, and in other developing
regions. As such, the incumbent state-owned utility typi-
cally continues to play a key role in the sector, but because
of inefficiencies and inadequate investment resources, IPPs
are gradually being introduced. These hybrid power
markets are giving rise to new challenges that need explicit
attention if private investment is to be accelerated.
In addition to introducing legislation allowing for pri-
vate power generation, the eight countries examined in this
chapter also have established independent regulators. In
Kenya the presence of the regulator, together with the adop-
tion of International Competition in Bidding (ICB) prac-
tices, helped reduce PPA charges radically between the first
set of IPPs and the second. A similar trend may be observed
in Senegal, where the first IPP (GTi Dakar) was not overseen
by the independent regulatory body and the second (Kou-
noune I) was (Regulatory Commission of Senegal 2010; IFC
2010b). Kenya’s Energy Regulatory Commission (ERC) has
also been instrumental in helping set tariffs and manage the
overall interface between private and public sectors. In
Uganda sponsors note the benefits of having the regulator
involved from project inception, particularly in helping to
increase overall transparency, especially in the case of
Bujagali. An ERC staff member affirms that the commis-
sion’s “presence has helped to focus minds on the require-
ments for setting up power supply projects so that investors
coming in are clear of what is expected of them from the
beginning and hence align their bids to these require-
ments. As a result, we have increasing numbers of investors
applying to set up IPP projects” (ERC 2010a).
In Côte d’Ivoire, Ghana, Nigeria, and Tanzania, however,
regulatory agencies have come into force only after IPPs
have been negotiated, and as yet the agencies have had little
impact on new investment. In general, the mere presence of
a regulator does not appear to be a defining factor in attract-
ing IPPs. An independent regulator may have positive, neg-
ative, or no impact on IPP outcomes. If, however, regulatory
governance is transparent, fair, and accountable, and if reg-
ulatory decisions are credible and predictable, there is
greater potential for positive outcomes for both the host
country and investor alike. Evidence also suggests that effec-
tive regulatory oversight may lead to a reduction in the
stated capital costs of projects and improved efficiency for
selectively bid projects (Phadke 2007; Eberhard and
Shkaratan 2010).7
In Nigeria and Tanzania efforts have been made to
exploit stranded gas as part of the IPP program.8 In Nigeria
a reduction in gas flaring is central to the push for gas-fired
power, while in Tanzania the IPP program commercialized
previously stranded (although not flared) gas through
Songas and Mtwara (a small private concession in the south
of the country). Although these two countries have a dis-
tinct set of challenges,9 in general the larger policy of involv-
ing stranded gas has insulated projects from intense public
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
scrutiny, with project sponsors and policy makers alike able
to point to the benefits of the commercialized gas and the
reduction in fuel imports.
Finally, DFIs are behind many countries’ power sector
policy frameworks; notably the World Bank has had a hand
in nearly all power sector reform programs in Africa. DFIs
have been particularly instrumental in advancing private
sector participation in generation. And as many of those
same institutions began reconsidering publicly funded
infrastructure investments at the end of the 1990s, coun-
tries’ policies have followed—from state to market and back
again, albeit with some changes to accommodate the more
hybrid market that has emerged.
Linking planning, procurement, and contracting
Intricately connected to sound policy frameworks are
coherent power sector plans. Ideally, planning, procure-
ment, and contracting follow from sound policy frame-
works and include a number of core components: setting a
reliability standard for energy security, completion of
detailed supply and demand forecasts, development of a
least-cost plan with alternative scenarios, clarification of
how new generation production will be split between the
private and public sectors, and identification of requisite
bidding and procurement processes for new projects. One
important aspect of coherent power sector planning is vest-
ing planning and procurement in one empowered agency to
ensure that implementation takes place with minimal
mishaps (Eberhard and Malgas forthcoming).
While that planning and procurement arrangement may
be ideal, the reality is often quite different. As one stakeholder
at Ghana’s Public Utilities Regulatory Commission (PURC)
notes about the recent past, “A crisis arises, and everybody
panics; anybody who comes in [to propose generation] is
listened to” (PURC 2010). Indeed, in the eight countries
examined here, there have been several noteworthy plan-
ning mishaps that subsequently affected procurement and
contracting. In some cases, demand and supply were not
being accurately forecast partly because of weather condi-
tions such as extended droughts, which in turn necessitated
fast-tracking IPPs—that is, plans for IPPs were sped
through to meet immediate power shortages. The first
two plants in Kenya (Westmont and Iberafrica), the first
plant in Nigeria (AES Barge), and Ghana’s Takoradi II IPP
were all negotiated during drought conditions. Although
both the Westmont and Iberafrica plants came online
within 11 months, they were later investigated for suspicious
bidding practices and what were perceived as unnecessarily
375
 expensive charges. Furthermore, public stakeholders’
unwillingness to make concessions over a tariff resulted in
Westmont not securing a second PPA. The AES Barge proj-
ect in Nigeria took nearly two years to come online because
of a renegotiation of the PPA, despite the project being
fast-tracked. In Ghana, no agreement was reached over the
second phase of the Takoradi II project.
Inability to estimate demand and supply accurately or to
set a clear reliability standard has also led to several cases in
which emergency power units needed to be leased for one to
two years with the purpose of plugging a short-term crisis.
In 2009 approximately 750 MW of emergency power was in
operation in Sub-Saharan Africa (Foster and Briceno-
Garmendia 2010). The governments of Ghana, Kenya,
Tanzania, and Uganda all have been forced to address
drought and black-outs in recent years, often turning to
emergency power suppliers (see box 21.1 for an example).
Kenya harnessed 100 MW of emergency power twice: in
1999, 2000, and 2001, and again in 2006 (supplied by Aggreko,
Cummins, and Deutz in the first instance and Aggreko alone
in the second). In 2007 Aggreko’s contract was extended for
an additional two years, and in 2009 Aggreko was selected to
provide a further 140 MW for a total of 250 MW of emer-
gency power.10 In Ghana emergency power was instrumental
in reducing the impact of the 1998 drought, but as drought
conditions reversed, the state failed to honor its contracts
with SIIF Accra; as of 2007 those contracts were still in dis-
pute. The cost of emergency power, at approximately 30 to
40 cents a kilowatt hour (kWh), is high (IFC 2005). Tanzania,
for example, has estimated that it saves around $1 for every
kWh of outage averted (or about 5 to 10 times the ordinary
cost of generating electricity).11
Competitive International Bidding Practices.
Although it is easy, in hindsight, to accuse IPP stakehold-
ers of acting imprudently in the face of emergencies, the
actual conditions of load-shedding and shortages in the
power sector appear to have provided few alternatives. Two
studies highlight the importance of the international com-
petition in bidding practices in reducing up to 60 percent
of the capital cost of plants (Deloitte Touche Tohmatsu
Emerging Markets Ltd. and Advanced Engineering Associ-
ates International 2003; Phadke 2007). There is also evi-
dence for selective bidding proving effective in certain
instances, provided there is regulatory scrutiny. Of the 21
IPPs examined here, ICBs are known to have been con-
ducted for nine. Six of the projects in East Africa (OrPower
4, Tsavo, Rabai, Songas, Namanve, and Bujagali) con-
formed to these bidding practices, as did three projects in
376 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
Box 21.1 Poor Planning Led to Significant Delays
for Tanzania’s Projects
The quick and obscure process through which IPTL
(Tanzania) was negotiated resulted in the highest-
profile IPP mishap on the continent to date. The
project’s total investment costs were estimated at
$150 million ($163 million including fuel conver-
sion), an amount the government and the World
Bank later argued was inflated by 40 percent. This
dispute led to a three-year arbitration process during
which another project in Tanzania, the Songo Songo
gas-to-electricity project, was also put on hold,
mainly through pressure from the World Bank, its
largest donor, because of alleged corruption.
Although the IPTL arbitration reduced its estimated
costs it was still well above the price the government
sought to pay. In addition, since both IPPs were
unavailable until 2002 and 2004 respectively, the state
had to negotiate with another company, Rich-
mond/Dowans, for emergency power, a contract that
came with its own inflated costs and controversies.
Furthermore, because of the delays, Songo Songo
accumulated $100 million in interest charges on
owner’s equity.
In this project, critical planning elements are miss-
ing, namely, a clear reliability standard, an accurate
demand and supply forecast, a detailed plan for pri-
vately powered and publicly powered generation, and,
most important, timely initiation of procurement and
effective conclusion of contracts.
Source: Authors.
West Africa (Azito, GTi Dakar, and Kounoune I). Of the
projects examined here that have faced renegotiation, four
were not bid through an ICB (IPTL, Iberafrica, AES Barge,
and Okpai); the two exceptions were Songas and OrPower
4. Absence of regulatory scrutiny is also noteworthy in
each of these four projects. Furthermore, Westmont,
which was selectively bid, quit the country after its first
seven-year PPA expired. Other non-ICB projects have also
encountered difficulties. For example, Ghana’s Takoradi II
has not been able to raise financing for the second phase of
the plant. Although reasons for these stumbling blocks
may be traced well beyond the presence or absence of an
ICB, the correlation is nonetheless revealing.
Furthermore, the success of the ICB process is intricately
linked to the number of bids received, with more bidding
driving down prices,12 as evidenced by IPPs in North Africa,
where the number of bids submitted to ICBs has been gen-
erally double to triple those submitted to ICBs in East and
West Africa. That said, the time and cost required to com-
plete an international competitive bid should not be under-
estimated. In Sub-Saharan Africa, drought-related energy
crises are often cited as the reason why ICBs have not been
pursued. In Togo, for example, the drought conditions of
2006 prompted a move to discontinue a rehabilitate-own-
transfer contract with Electro Togo to manage the Centrale
thermique de Lomé. Rather than launching an ICB, the gov-
ernment chose to negotiate directly with an existing player,
Contour Global, which was already in discussions with the
utility. In this case, time and project familiarity proved more
important than complying with international bidding prac-
tices, which risked extending the project development time-
line (Ministry of Energy of Togo 2010).13
The reason for these planning and contracting mishaps
seems to lie in the changing nature of power markets
across Africa and other developing regions. In addition, as
power markets continue to include private sector partici-
pation, it is not always clear who has responsibility for
maintaining security of supply. Often, the planning func-
tion is shifted from the state-owned utility company to
the energy ministry, which does not have the capacity,
resources, or experience to undertake detailed power sector
planning. Alternatively, planning is contracted to consult-
ants who produce master plans that quickly become
outdated as global equipment, fuel costs, and other key
parameters change (Eberhard and Malgas forthcoming).
Absence of regulatory control and institutional capacity is
evident in more areas than just planning. In many
instances, there are no clear criteria for allocating power
plant building opportunities to either the incumbent
state-owned utility or the private sector.
Often, it is not clear whether plans are merely indicative,
whether unsolicited proposals may be considered, or
whether plans have legal force in determining which plants
the regulator may license. Too often, plans do not translate
into timely initiation of competitive bid processes for new
plants. Similarly, there is often insufficient capacity to nego-
tiate with winning bidders or to conclude sustainable con-
tracts. Transaction advisers may be appointed, but often
there is little continuity and the overall policy framework is
lacking information about security packages or credit
enhancement measures being offered by government.
Despite past planning deficiencies and forced reliance on
emergency power generation, Kenya provides an interesting
recent example in which progress is being made in dealing
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
with these planning, procurement, and contracting chal-
lenges (box 21.2).
Abundant, low-cost fuel and
secure fuel contracts
The availability of competitively priced fuel supplies has
also emerged as a key factor in how IPPs are perceived and
ultimately whether there is public appetite for such projects,
in large part because fuel is generally a pass-through cost to
the utility and, in many cases, also to the final consumer.
Thus, if the IPP uses a different fuel than other plants, and
if that fuel is more expensive, there is greater potential for
disputes surrounding the new project.
In several Sub-Saharan African countries—Ghana, Kenya,
Senegal, Tanzania, and Togo—low-priced hydropower was
the dominant fuel source for power production, but IPPs
were thermal powered, using a combination of imported fuel
Box 21.2 Planning, Procurement, and Contracting
Challenges in Kenya
Recognizing that it does not have the internal capacity,
resources, or planning tools to develop detailed and
up-to-date electricity plans, Kenya’s Energy Regulatory
Commission (ERC) has delegated this function to the
Kenya Power and Light Company (KPLC), guided by a
government committee chaired by the ERC. KPLC
ceased being an electricity-generating company in
1997, and thus has a neutral stance between the state
utility, Kenya Electricity Generating Company
(KenGen), on one hand and private IPPs on the
other. KPLC has also been assigned responsibility
for managing the procurement and contracting
process for IPPs. As described by one project sponsor
in Kenya, commenting in May 2010, “Kenya has an IPP
structure that is working. It has a track record. It can
structure new projects based on experience gained
from previous projects. And it has a very capable set of
teams working in KPLC, the Ministry of Energy/
Finance and KenGen. It understands project finance
and is not surprised when a developer requests a com-
fort letter, as one example.” The fruits of this approach
are evident. Kenya is about to add three new IPPs to its
existing five, maintaining its lead position in Sub-
Saharan Africa in terms of IPP investments.a
a. Kenya also has been cited by public stakeholders in Ghana
and Tanzania as having processes that should be emulated
(PURC 2010; EWURA 2010).
377
 oil and domestically procured natural gas. These same coun-
tries witnessed a series of debilitating droughts during the
1990s and 2000s, during which existing hydropower infra-
structure proved insufficient, and thermal power, provided
almost entirely by IPPs, was increasingly integrated into the
fuel supply mix (from 10 percent to 60 percent in Tanzania,
for example), forcing up the price of power. Although more
thermal power may be required, the public perception has
been that IPPs drive up the price of power. OrPower 4,
Kenya’s geothermal IPP, deserves special mention in the con-
text of hydro-dominant systems that have been diversified
with (largely imported) fuel oil (box 21.3).
The number of IPPs in Sub-Saharan Africa with secure,
low-cost fuel sources is still relatively few. In Tanzania,
although natural gas from the Songo Songo field is cheaper
than the imported fuel oil currently powering IPTL,14 dis-
putes continue to surround its use, and IPTL’s diesel units
have yet to be converted to gas. In Ghana the Sunon Asogli
Power Plant is completed but awaits fuel. Ghana’s Volta
River Authority (VRA), the state-owned generator, is a
foundation customer for the West African Gas Pipeline
(WAGP), from which it has been receiving gas since Febru-
ary 2010, but the recently completed Shenzhen IPP will
receive gas only when the pipeline is pressurized and higher
volumes of gas flow. In short, an IPP’s access to abundant,
low-cost fuel is not only a matter of whether such fuel is
available in a country, but whether the supply of such fuel is
Box 21.3 Petroleum-Fired Versus Natural
Gas-Fired IPPs
As indicated in a press release from Ormat (2009), “At
a price of US$29 per barrel of petroleum crude oil, [a]
48 MW geothermal plant is cheaper to operate than a
heavy fuel oil fired plant. This means that at the cur-
rent oil prices still above US$40 per barrel, OrPower 4
is providing cheaper electricity to the national grid
than any existing oil fired plants in Kenya.” In con-
trast, Nigeria has relied entirely on domestically pro-
cured natural gas to fuel its IPPs, and gas is the incum-
bent fuel. Until recently, although a series of issues
affected project outcomes, most notably the invest-
ment climate and bidding procedures, fuel had not
been an issue; however, civil unrest in the Niger Delta
led to a disruption in the fuel supply in 2007–09,
albeit with improved conditions as of 2010.
Source: Authors.
378 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
guaranteed through contracts well into the future. Fuel
sourcing and other country-level elements that contribute
to successful IPPs are summarized in table 21.2.
PROJECT-LEVEL FACTORS CONTRIBUTING
TO SUCCESSFUL IPPS
Investors in IPPs in Sub-Saharan Africa must navigate
changing investment climates, national policies, and plan-
ning frameworks. Starting with an evaluation of equity
arrangements, this section examines trends in investor
behavior and how investors secured revenue to service debt
and reward equity, particularly in the face of exogenous
stresses.
Favorable equity arrangements
In assessing the project-level factors contributing to the suc-
cess of IPPs, it is useful to ask several questions. Does the
presence of local equity shareholders make a difference in
project outcomes? Are projects with such participation less
likely to face pressure from host-country governments to
change their contract terms? How does a firm’s previous
experience with a country play out in making or breaking
deals? Does the presence of DFI or firms with international
development experience, such as Industrial Promotion Ser-
vices (IPS), Globeleq, and Aldwych International, affect the
likelihood of an IPP’s success? Table 21.3 shows the country
origin of Sub-Saharan African IPP sponsors, along with
their respective equity share, whether the project faced a
change in contract terms, and whether there was turnover of
the majority equity partner.
Unlike in China and Malaysia, where local investors in
IPPs abound, foreign firms have long been dominant play-
ers in Sub-Saharan African IPPs (Woodhouse 2005). Given
the limited availability of capital in Sub-Saharan African
countries in the sample, this is not surprising. Foreign
involvement, however, raises the issue of foreign exchange
exposure.15 Only 3 of the 21 projects examined here have
local majority stakeholders: Nigeria’s Okpai, Afam VI, and
Aba Integrated. In two of these cases, however, the stake-
holder was either the national utility or the Nigerian
National Petroleum Company (NNPC).
Several observers (for example, Hoskote 1995; Wood-
house 2005) have cited local participation in IPPs as a
possible means of reducing risk. Ten of the 21 projects
have local equity participation—namely, Sunon Asogli,
Iberafrica,16 IPTL, Songas, Takoradi II, AES Barge, Okpai,
Afam VI, Aba Integrated, and Bujagali. As noted, six of these
 Table 21.2 Elements within Purview of Host Governments Contributing to Successful IPPs
Element
Favorable investment climate
Clear policy framework
Clear, consistent, and fair regulatory oversight
Coherent power sector planning linked to procurement
and contracting
Abundant, low- cost fuel and secure contracts
Source: Authors.
projects have encountered some form of change to their
contract, and in four of the six, either the state utility or
another government entity held an equity share, indicating
that the existence of a local partner might not be critical in
preserving the original financing balance.
As for renegotiating contract terms, it is unclear
whether having a local partner makes a difference. Kenya’s
Westmont and Iberafrica, for example, were negotiated at
the same time under similar policy frameworks, although
Iberafrica had local participation and Westmont did not.
Iberafrica first voluntarily reduced its tariff and then went
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
Details
■ Stable macroeconomic policies
■ Legal system that allows contracts to be enforced, laws to be upheld,
and arbitration
■ Good repayment record and investment-grade credit rating
■ Requires less (costly) risk mitigation techniques to be employed,
translating into lower cost of capital, lower project costs, and more
competitive prices
■ Potentially more than one investment opportunity
■ Policy framework enshrined in legislation
■ Framework clearly specifies market structure and roles and terms for
private and public sector investments (generally for single-buyer model,
as wholesale competition does not yet exist in African context)
■ Framework is led and implemented by reform-minded “champions,”
concerned with long-term power sector conditions
■ Oversight improves general performance of private and public
sector assets
■ Transparent and predictable licensing and tariff framework improves
investor confidence
■ Cost-reflective tariffs ensure revenue sufficiency
■ Consumers protected
■ Energy security standard in place; planning roles and functions
are clear
■ Power planning vested with lead, appropriate (skilled, resourced, and
empowered) agency
■ Power sector planning takes into consideration the hybrid market
(public and private stakeholders and their respective real costs of
capital) and fairly allocates new building opportunities among
stakeholders
■ Planning has built-in contingencies to avoid emergency power plants
or blackouts
■ Responsibility for procurement is clearly allocated; plans are linked
to procurement and bids are initiated in time
■ Procurement process is transparent and competition ultimately drives
down prices
■ Capacity is built to contract effectively
■ Cost-competitive with other fuels
■ Contract safeguards affordable and reliable; fuel supply sufficient for
duration of contract
on to negotiate a second 15-year PPA, in contrast to West-
mont, which stopped production after failing to come to
an agreement on a second PPA. Although the presence of a
local partner may have helped Iberafrica create a long-
term solution, with just one example, the evidence is not
conclusive. Togo’s Centrale thermique de Lomé, due to be
online in July 2010, may provide a more sustainable
method for balancing investment and development out-
comes. In this project, 25 percent of the project equity
must be sold to local investors within the first five years
(Ministry of Energy of Togo 2010).
379
380 Table 21.3 Equity Participation in IPPs in Sub-Saharan Africa
Majority equity
Change in partner turnover
Project Equity partners (country and percent of equity held) contract terms (number of times)
Westmont (Kenya) Westmont (Malaysia, 100%); has sought to sell plant since 2004 — —
Iberafrica (Kenya) Union Fenosa (Spain, 80%), KPLC Pension Fund (Kenya, 20%) since 1997 yes 0
OrPower 4 (Kenya) Ormat (United States/Israel, 100%) since 1998 yes 0
Tsavo (Kenya) Cinergy (United States) and IPS (international) jointly owned 49.9%, Cinergy sold to Duke Energy no 1
(United States) in 2005; CDC/Globeleq (United Kingdom, 30%), Wartsila (Finland, 15%), IFC
(international, 5%) have retained remaining shares since 2000
Rabai (Kenya) Aldywch-International (Netherlands, 34%), BWSC (Denmark, but owned by Mitsui of Japan, 25.5%), no 0
FMO (Netherlands, 20%), IFU (Denmark, 20%)
IPTL (Tanzania) Mechmar (Malaysia, 70%), VIP (Tanzania, 30% in kind); both have sought to sell shares, and actual equity yes —
contribution is currently disputed
Songas (Tanzania) TransCanada sold majority shares to AES (United States) in 1999, and AES sold majority shares to yes 2
Globeleq (United Kingdom) in 2003.a All preferred equity shares were converted into “loan notes”
in June 2009, only common shares remain
Bujagali (Uganda) Sithe Global (United States, 58%), IPS-AKFED (32%), government of Uganda (10%) no 0
Namanve (Uganda) Jacobsen (Norway, 100%) no 0
CIPREL (Côte d’Ivoire) SAUR International (88%). This is a joint venture between French SAUR Group owned by Bouygues no 1
(65%), and Electricité de France (35%). The remaining 12% was owned by a combination of BOAD
(West African Bank for Development); the Investment and Promotions Company for Economic
Cooperation, a subsidiary of AFD; and IFC. In 2005 98% of the shares were sold to Bouygues
(France); BOAD retained 2%.
Azito (Côte d’Ivoire) Cinergy (65.7%; joint venture between Swiss ABB and Globeleq), CDC/Globeleq (11%), and IPS-AKFED (23%) no 1
Takoradi II (Ghana) CMS (United States, 90%), VRA (Ghana, 10%); CMS sold shares to TAQA (UAE, 90%) in 2007 yes 1
Sunon Asogli (Ghana) Shenzhen (China), Togbe Afede XIV (Ghana/local strategic investor) no 0
Bui hydro (Ghana) Sinohydro (China) no 0
AES Barge (Nigeria) Enron (United States, 100%); sold to AES (United States, 95%) and YFP (Nigeria, 5%) in 2000 yes 1
Okpai (Nigeria) Nigerian National Petroleum Corporation (Nigeria, 60%), Nigerian Agip Oil Company (Italy, 20%), yes 0
and Phillips Oil Company (United States, 20%) have maintained equity since 2001
Afam VI (Nigeria) Nigerian National Petroleum Corporation (Nigeria, 55%), Shell (United Kingdom/Netherlands, 30%), no 0
Elf (Total) (France, 10%), Agip (Italy, 5%)
Aba Integrated (Nigeria) Geometric Power Limited (Nigeria) no 0
GTi Dakar (Senegal) GE Capital Structured Finance Group (SFG) (USA), Edison (Italy), IFC no 0
Kounoune (Senegal) Mitsubishi (Japan), Matelec S.A.L (Lebanon) no 0
Centrale thermique Contour Global (United States, 80%) IFC (20%) no 0
de Lomé (Togo)

Source: Authors.
Note: — = Pending developments.
a. Turnover within the Songas project has been complex: Ocelot (Canada), TransCanada (Canada), Tanzania Petroleum Development Corporation, TPDC (Tanzania), TANESCO (Tanzania),
Tanzania Development Finance Company, TDFL (Tanzania, sponsored by European Investment Bank), IFC (multilateral), DEG (Germany), and CDC (United Kingdom) were shareholders
by 1996, with TransCanada the majority shareholder. IFC and DEG sold their shares to CDC in 1997/98; TransCanada sold its shares to AES (United States) in 1999; Ocelot/PanOcean sold
shares to AES in 2001; and AES sold majority shares to Globeleq (United Kingdom) and FMO (the Netherlands) in 2003. After the AES sale, equity shares and associated financial commit-
ments in Songas were as follows: Globeleq, $33.8 million (56 percent); FMO, $14.6 million (24 percent); TDFL, $4 million (7 percent); CDC, $3.6 million (6 percent); TPDC, $3 million
(5 percent); and TANESCO, $1 million (2 percent). The amount given for Globeleq does not reflect the additional $50 million that the company committed for the expansion of the project.
Origins, Experience, and Mandates of Partners. From
a global perspective, IPP investments during the 1990s were
led by a host of American and European investors who saw
returns in their home markets diminishing. A wave of
investors originating from developing countries, however,
particularly from Malaysia, was also present. Although it
would be inaccurate to say that investors based in develop-
ing countries overlooked the risk involved in African coun-
tries (or did not ultimately charge higher returns), they may
have had a greater willingness to consider IPP investments
in Sub-Saharan Africa in the first place.
While the number of developing-country-based investors
in IPPs appears to be growing, three such firms are trying to
sell their shares (Mechmar, VIP, and Westmont). Thus, the
home country of the firm does not mean that project equity
is permanent, or that firms based in developing countries are
best positioned to service debt and reward equity.
An aspect more important than the nationality of the
investing firm appears to be a firm’s experience and mandate.
Across the pool of IPPs examined in table 21.3, several firms
were actively involved in the country before making an IPP
investment. Union Fenosa, for example, the parent company
of Iberafrica, had previous experience in Kenya through an
information-technology contract. IPS, a major shareholder in
Tsavo (Kenya), Azito (Côte d’Ivoire), and Bujagali (Uganda),
has operated in Kenya since 1963 and in Côte d’Ivoire since
1965. For certain projects, it may be argued that long-term
relationships, particularly those with strong local manage-
ment, appear to have contributed to the staying power of
firms and often the rebalancing of contract terms.
In addition to the terms of the deal, the mandate of the
investing firm appears to play a central role in firms’ deci-
sions about whether to invest in IPP projects. Although
Globeleq, IPS, and Aldwych, for example, are commercial
entities, they emerged from agencies with a strong commit-
ment to social and economic development. Until recently,
the two firms that were increasing their stakes were Glo-
beleq and IPS. Globeleq holds a 43 percent share in Côte
d’Ivoire’s Azito, 30 percent equity in Kenya’s Tsavo, and
56 percent in Tanzania’s Songas. IPS holds a 23 percent
share in Azito, and together with Duke Energy, a 49.9 percent
share in Tsavo. IPS is also leading development of
Uganda’s Bujagali project and is a 35 percent shareholder
in equity. Although a smaller player than either Globeleq
or IPS, Aldwych International has also made significant
inroads via Rabai in Kenya and is evaluating further
expansion.
With the exception of Tanzania’s Songas, none of the
Sub-Saharan African IPP projects with involvement of firms
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
with development origins examined here has had any
changes in contract terms, which may signal that they were
better balanced from inception in terms of investors and
had a better ability to withstand public pressure. Further-
more, in terms of the Songas change, although the
$103 million buy-down of the allowance for funds used
during construction resulted in a reduction in the capac-
ity charge, the firm received full payment upon the buy-
down—a different case from many of the contract changes
cited earlier. Alongside companies with development ori-
gins, foreign investors remain directly involved in IPP
investments. Indeed, DFI equity shares have increased in
recent years for at least five IPPs. With the exception of
Songas, no project with DFI involvement has seen any con-
tract changes. Box 21.4 provides further insights on IPP
projects concerned with their developmental impacts.
Equity Turnover. Of the 46 original equity partners in
the 21 projects considered here, 7 have exited (from 5 pro-
jects). This ratio, however, tells only part of the story. First,
as previously indicated, shareholders in Westmont (Kenya)
and IPTL (Tanzania) have been trying for several years to
Box 21.4 Two IPPs with Development and
Investment Mandates
Development is perhaps more important for Indus-
trial Promotion Services (IPS) than for any other firm
investing in IPPs in Sub-Saharan Africa. As a rule, IPS
invests only in projects with a substantial develop-
mental impact and a reasonable internal rate of
return. For the Tsavo project, that rate is 17–18 per-
cent, and for Bujagali, 19 percent—considerably lower
than typical internal rates of return for IPP projects in
Sub-Saharan Africa (IPS 2010).
Aldwych International, on the other hand, requires
that its project investments both make commercial
sense and serve a clear developmental function for the
country and local community. The company’s experi-
ence in the Rabai in Kenya helps to illustrate this
point. The firm has faced three major hurdles as part
of its involvement in the project—legal trouble during
the tendering process, national postelection civil
unrest in late 2007, and the global financial crisis.
Despite the challenges, the project reached financial
close successfully in 2008; it even won the Project
Finance International’s deal of the year award.
Source: Authors.
381
 sell their assets. In the case of IPTL, Mechmar, the lead
shareholder, has indicated that the sale has been driven by
an arbitration settlement that hurt its equity partners.
VIP, the minority shareholder, cites oppression by the
majority shareholder, fraud by Mechmar in inflating the
IPTL capital cost, and failure by Mechmar to pay its
equity contribution (that is, the project was 100 percent
debt financed) as reasons behind its desire to sell its stake.
For this and other projects, investors repeatedly say that
their motivation to sell stakes in IPPs is driven primarily
by changing circumstances in home markets or corporate
strategy—that is, the desire to sell has little to do with
host-country actions and reactions or with poor invest-
ment outcomes, namely, the ability to service debt ade-
quately and reward equity.
While investment outcomes may be partially motivating
sales, turnover does not in and of itself appear to be chal-
lenging the long-term sustainability of contracts, since in
nearly all cases sellers have found willing buyers to take over
the original or recently renegotiated PPAs. The two excep-
tions, again, are Westmont, in which the first PPA expired
and was shrouded in controversy, and IPTL, which has been
embroiled in lawsuits. Under such circumstances, it may
therefore be understandable that the plants have not
attracted buyers. One stakeholder went so far as to assert
that “[equity turnover is a] healthy factor in a maturing
market. It is a good sign when investors come and go—not
a bad or threatening thing.” The return of the government
as a shareholder, as planned in the case of Tanzania’s IPTL,
would, however, signal that some markets are less mature
than initially expected.
Debt arrangements: Global and local
With debt financing often representing more than 70 percent
of total project costs, competitively priced financing has
emerged as a key factor in successful projects. How and where
to get this low-cost financing is a challenge, but possible
approaches in the case of Sub-Saharan Africa are DFIs,
credit enhancements, and flexibility in repayment terms and
conditions, including possible refinancing.17 The goal for
sustainable financing should be that the risk premium
demanded by financiers or capped by the off-taker matches
the actual country risk and that project risks and risk pre-
miums are not inflated.
While there is no uniform pattern in the debt financ-
ing for the projects considered here, observation of sev-
eral trends in terms of how investors handled costs may
contribute to the success of other projects. Although
382 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
nonrecourse project financing is the norm for privately
financed electric power plants in developing regions, the
21 projects examined here represent several notable excep-
tions: Nigeria’s Okpai and Afam VI, which were 100 percent
financed by the balance sheet of equity partners, and the
second phase of Songas, which was largely refinanced by a
World Bank loan in 2009 (Globeleq 2010a). Until recently,
Westmont, Iberafrica, and OrPower 4 were all financed
entirely with the balance sheets of their sponsors. For West-
mont and Iberafrica, the reason cited for this arrangement
was insufficient time to arrange project finance, because
plants had to be brought online within 11 months. For
OrPower4, the reason was linked to lack of a security pack-
age, which was not finalized until 2006.
The Impact of DFIs on Projects. With limited appetite
for IPPs among commercial banks, DFIs provide a sub-
stantial amoung of credit to projects. Indeed, DFIs such as
the World Bank, International Finance Corporation (IFC),
European Investment Bank, DEG, FMO, African Develop-
ment Bank, PROPARCO, the Emerging Africa Infrastruc-
ture Fund, and European Financing Partners, KfW, and
Agence Française de Développement have participated in
nearly every IPP. Beyond these institutions’ long history of
activity in Sub-Saharan Africa, their involvement is attrib-
utable to real and perceived risks by private investors,
which keep them from filling the financing gap, and to the
DFIs’ interest in participating in the broader mandate of
power sector reform.
Although projects with DFI funding tended to take
longer to reach financial closure than those financed
through private sources, project sponsors in Kenya say that
multilateral and bilateral development institutions helped
them maintain contracts and resist renegotiation in the face
of external challenges such as droughts, when developers
were pressured to reduce tariffs. A particularly revealing
contrast is between the two Kenyan plants, OrPower 4 and
Tsavo. Although the two projects were negotiated under the
same policy framework, the former initially had no multi-
lateral involvement in either its equity or debt,18 whereas the
IFC arranged all the debt for Tsavo and took a 5 percent
equity stake. Tsavo has since resisted KPLC pressure to
reduce its tariffs. OrPower 4, on the other hand, ultimately
reduced its tariff for the second phase of the plant. Tanza-
nia’s Songas project, for which the World Bank and Euro-
pean Investment Bank financed all project debt, also
deserves special mention here. The project took almost a
decade to reach financial closure; the World Bank played an
instrumental role in, among other things, pressuring the
IPTL arbitration, which ultimately led to more balanced
contract terms.
Locally Sourced Finance. While observers argue that
local financing is key to increasing sustainable foreign
investment, capital markets in many Sub-Saharan African
countries are underdeveloped and illiquid, hence unable
to provide financing for all projects. Three exceptions in
the project sample are Kounoune I of Senegal, Geometric
of Nigeria, and Namanve of Uganda. The Kounoune I
project received financing from both the West African
Bank for Development (BOAD), based in Togo, and the
Banking Company of West Africa (CBAO), a private bank
based in Senegal. Funding for Geometric has been pro-
vided by Diamond Bank and Stanbic IBTC Bank Plc, both
of Nigeria. Similarly, for Uganda’s Namanve project, financ-
ing was provided by Standard Chartered Bank of the United
Kingdom through Stanbic Bank of Uganda. By contrast, in
North Africa, all of the €213 million debt of Tahaddart, a
384 megawatt combined-cycle gas turbine plant in
Morocco, was financed by local banks. Local financing was
aided by a number of factors, including the state utility’s
prominent role in the plant (it holds nearly 50 percent of
total equity) and the fact that Morocco’s commercial banks
have a significant degree of state involvement. With or with-
out state involvement, however, no other country in Sub-
Saharan Africa has yet been able to manage to arrange this
level and depth of local financing for IPPs.
One main drawback for IPPs without local finance is the
impact of macroeconomic shocks and local currency deval-
uation. Since the late 1990s, Ghana, Kenya, and Tanzania
have experienced substantial depreciation, with their cur-
rencies losing more than 100 percent, 200 percent, and
400 percent of their value against the U.S. dollar, respec-
tively. Inevitably, these currency shifts have led to pressure
to reduce capacity charges and to countries reconsidering
IPP development.
Securing revenue: The PPA
All 21 of the projects examined here had a long-term power
purchase agreement with the incumbent state-owned utility
to ensure a market for the power produced and to secure
revenue flows for debt and equity providers. In addition to
stating who would buy the power, the PPAs detailed how
much power capacity would be available and how much
buyers would be charged. Provisions concerning fuel meter-
ing, interconnection, insurance, force majeure, transfer, ter-
mination, change of legal provisions, and refinancing were
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
generally also addressed. Nearly all the contracts specified
some form of international dispute resolution and mini-
mum power availability.
For each project, sponsors negotiated or were granted
PPAs in U.S. dollars, thereby reducing their exposure to
currency devaluation. Over time, the bulk of project
contracts have been upheld, but not without changes.
Box 21.5 reviews PPA renegotiations in Kenya, Nigeria,
and Tanzania.
Credit enhancements and security arrangements
The underlying credit risk of IPP projects in Sub-Saharan
Africa has been dealt with largely through a suite of credit
enhancements, namely, guarantees, insurance, and cash (the
latter of which took the form of escrow accounts, liquidity
facilities, and letters of credit of varying amounts and
tenures). The Tsavo project in Kenya, for example, has an
escrow account equivalent to one month’s capacity charge
and a standby letter of credit from KPLC covering three
months’ billing (approximately $12 million). At least 12 of
the 21 projects examined in this chapter had some form of
cash security arrangement, with typical terms of between
one and four months’ capacity charge in reserve.
Not surprisingly, the number of credit enhancements
appears to diminish as a country’s risk profile improves.
There are, however, noticeable exceptions such as the first
wave of IPPs in Kenya (Westmont and Iberafrica), for which
risk appears to be entirely reflected in the higher capacity
payments negotiated. That said, corruption was also alleged
in both these plants.19 Thus, the “security arrangement”
may lie not in a letter of credit but in an informal agreement
among sponsors.
Of the many different credit enhancements used in
connection with IPPs, sovereign guarantees are most
common. Such guarantees are known to have been
extended for at least nine of the pool of 21 projects:
Tanzania’s IPTL, Nigeria’s AES Barge, Côte d’Ivoire’s
Azito, Ghana’s Takoradi II (phase I), both GTi Dakar and
Kounoune I in Senegal, Togo’s Centrale thermique de Lomé,
and Bujagali and Namanve in Uganda. Several of the proj-
ects without guarantees—Tsavo and Rabai, for example—
were, however, given assurances by the government in the
form of comfort or support letters, through which political
risk is assumed. In the case of the Okpai plant in Nigeria,
security was extended in the form of the state-owned oil
company’s revenues. Thus, if the off-taker defaults, NNPC,
among the most liquid firms in the country, is liable. Partial
risk guarantees (PRGs) issued by the World Bank were used
383
 Box 21.5 The Experience of PPAs with Renegotiations in Kenya, Nigeria, and Tanzania

In Kenya’s first wave of IPPs, costs were inflated in part due to the short duration of contracts—only seven years. With
Iberafrica facing ongoing pressure to reduce its tariff, coupled with an interest in negotiating another contract, the
sponsor voluntarily reduced the capacity charge enshrined in the PPA. (Iberafrica’s second and third PPAs are for a
considerably longer time frame than its first, and tariffs have been reduced significantly.) Westmont, on the other
hand, did not negotiate a second contract after it failed to obtain the same terms—in particular, those related to capac-
ity charges—spelled out in its first PPA. Likewise, changes in Kenya’s OrPower 4 and Tanzania’s Songas PPA agree-
ments are related in part to the final amount of the capacity charge (as originally spelled out in the PPA).a
In the case of Nigeria’s AES Barge, initially sponsored by Enron, renegotiations in 1999–2000 brought about sev-
eral changes in the PPA, including a change in the fuel specifications (from liquid fuel to natural gas), which led to a
major reduction in the fuel charge for the off-taker. The current arbitration with AES Barge involves, among other
things, reconsideration of the availability-deficiency payment as well as the tax exemption. In each of the cases
reviewed here, the original terms of the PPA have been viewed as unsustainable for the host country and therefore
challenged.
The case of Tanzania’s IPTL is slightly different, however. Although the contract was considered initially unsus-
tainable due to the added capacity of Songas, the IPTL arbitration was prompted by what was deemed a breach in the
PPA, namely, the project sponsors’ substitution of medium-speed for slow-speed engines without passing on the cap-
ital cost savings to the utility, as per the PPA.b
a. It is, however, worth reiterating in this context that failure to agree on the security package and the capacity charge contributed
to delays in the development of OrPower 4’s additional 36 megawatts of production capacity.
b. Although this dispute was resolved, a subsequent and prevailing dispute relates to the level of actual equity in the project, which
in turn affects WACC and the allowed rate of return.

for two of the projects, Azito and Bujagali.20 In these
instances, the PRG covered all debt of the commercial
banks. In the case of default by the project company, the
PRG (backed by World Bank) would pay the commercial
lenders, and the World Bank would then claim repayment
from the government (World Bank 1997, 1999). For other
IPPs, political risk insurance was provided by the Overseas
Private Investment Corporation (OPIC), whereas guaran-
tees relating to currency inconvertibility, expropriation,
and political violence were issued by the World Bank’s
Multilateral Investment Guarantee Agency (MIGA).
In several cases, these credit enhancements have
improved the sustainability of projects and attracted or
assuaged lenders. The Bujagali project’s PRG, for example,
was instrumental in motivating and solidifying the involve-
ment of four commercial banks, which contributed a com-
bined $115 million at very competitive pricing. Some have
likened the PRG to a hammer effect, with the World Bank
guaranteeing what the government has already guaranteed
and thus making the government’s commitment twofold.
That said, PRGs are not appropriate for all Sub-Saharan
African IPPs because they are typically used for large proj-
ects in countries that are in an early stage of reform and
when commercial lenders are also present. Furthermore,
the government of the country must request a PRG; thus,
the project must be significant in the eyes of both the gov-
ernment and the World Bank. For projects without PRGs
security arrangements and credit enhancements are
similar, with the DFIs generally accepting the political risks,
such as in the Azito and Songas projects.
In Kenya, the only country among those in Sub-Saharan
Africa examined here to extend sovereign guarantees to
IPPs, stakeholders in Tsavo indicated that, without such a
guarantee, the presence of the IFC became critical, both to
help arrange debt and share in equity. In Ghana, lack of sov-
ereign guarantees has been cited as the main obstacle to
developing the second phase of the Takoradi II project.
Other credit enhancements have been used in abundance
in Kenya, including a suite of escrow facilities, which have
contributed to KPLC’s cash-strapped position. Although
government guarantees were recently debated by officials,
it looks as if the government will retain its no-guarantee
policy going forward, providing only letters of support to
IPPs. KPLC cites the absence of sovereign guarantees as
hampering its ability to raise private finance, while ERC
counters that IPPs have been introduced to help commer-
cialize the sector. Government guarantees work against
this goal, however.

384 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
 Notably, sovereign guarantees, political risk insurance
policies, and PRGs have not been invoked in any of the 21
projects examined here, including in those projects that
ultimately faced a change in the contract (AES Barge, IPTL,
OrPower 4, and Takoradi II). Recourse to international arbi-
tration has occurred only in the case of IPTL. Additionally,
it is helpful to reflect on the overall application of security
arrangements and credit enhancements. Although there is
some variation in the project sample, by and large the vari-
ation is limited, with the size of the project, the track record
of the regulatory regime (including its stability and credi-
bility), and the creditworthiness of the off-taker being the
main determinants. While there is resistance to government
guarantees on the part of some country stakeholders, proj-
ect developers and multilaterals typically support them. As
one World Bank (2010) official noted, “the first level of sup-
port has to come from the government.”
Furthermore, there has been very little evolution in the
credit arrangements used for IPPs in Sub-Saharan Africa.
All projects are supported by a PPA and their credit risk
largely carried by a government guarantee. In countries in
which the market for power is not developed, a PPA, along
with a government guarantee, is generally required, partic-
ularly where the off-taker is not creditworthy (IFC 2010a;
Rudo 2010a). This situation is quite different than that in
other developing regions. In Latin America, for example,
PRGs and other credit enhancements and security
arrangements are virtually nonexistent because power
markets are operational and local lenders are “in the
driver’s seat and generally very comfortable with local
developers and regulation.”21
Positive technical performance
Virtually all IPPs among the 21 considered here have
shown positive technical performance. Exceptions include
Nigeria’s plants (which have had problems with fuel sup-
ply) and, more recently, Kounoune I in Senegal (Nigeria
Electricity Regulatory Commission 2010; IFC 2010b). In
general, IPPs’ technical performance is superior to that of
state-owned plants. An argument for extending gas from
the West African Gas Pipeline to the Sunon Asogli Power
Plant in Ghana, for example, is being made in part because
Sunon’s technical performance is superior to that of the
Volta River Authority (PURC 2010). The Ministry of
Energy of Togo (2010), meanwhile, has indicated that Cen-
trale thermique de Lomé is expected to be more efficient
than its state-owned counterpart. In terms of availability,
IPPs in Kenya had an average availability of approximately
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
95 percent between 2004 and 2006, versus 60 percent for
KenGen’s thermal plants. When asked generally about his
most favorable IPP experience to date, one public stake-
holder simply indicated, “their technical performance”
(ERC 2010).
Strategic management and relationship building
Once 20-year contracts are in place, it would seem that the
deal is set and the revenue secured, with clear provisions to
ensure debt repayment and reward equity. Several other
interrelated actions deserve mention, however. One involves
relationship building, including with local partners and
communities. Another relates to how sponsors handle the
onset of stresses, including those presented by capacity
charges and refinancing.
Numerous IPP sponsors have adopted outreach pro-
grams to improve relations with local communities. In
Kenya, for example, Tsavo set up a $1 million community
development fund for the duration of the 20-year PPA,
from which grants of $50,000 each are disbursed each year
to benefit environmental and social activities in Kenya’s
Coast Region. Iberafrica maintains a social responsibility
program, and IPTL also is an active donor to its immedi-
ate community. In Ghana CMS’s social responsibility
involvement (before it sold its shares to TAQA in 2007)
included providing scholarships for secondary and tertiary
education and supporting medical clinics and the con-
struction of drainage systems. Bujagali also has a suite of
social outreach programs. Although the sums are not sig-
nificant, these programs, particularly when well adver-
tised, have the potential to win allies and counter the
stereotype of IPPs.
Another, perhaps more significant, action is how spon-
sors cope with stresses such as macroeconomic shocks, cur-
rency depreciation, and pressure from host governments to
reduce costs. Anecdotal evidence suggests that strategic
management helped put Kenya’s Iberafrica back on track, in
contrast to Westmont, where there is no evidence of such
action. Iberafrica, in fact, has had to cope with two major
stresses, drought and alleged corruption. According to
stakeholders at Iberafrica, the IPP voluntarily reduced its
capacity charge when KPLC was operating at a loss (due
in part to a drought-related recession), to show its sup-
port for the country and signal its interest in a second
contract. Iberafrica later secured a second contract, albeit
after even further reductions were negotiated and passed
by the electricity regulator.22 For Rabai, another project in
Kenya, sponsors chose to continue work on the project
385
 in the midst of countrywide protests following the elec-
tion in December 2007 (Aldwych International 2010a).
Project refinancing
A final area in which IPPs may yield greater balancing in
terms of development and investment outcomes is in the
refinancing of projects, evidence for which can be seen in
the Songas project in Tanzania. Possible refinancing in the
case of IPTL, with the government of Tanzania proposing
to buy the project’s outstanding debt equity, could also
lead to what may be perceived as a more balanced out-
come. A government buyback of IPTL would make a one-
time payment on behalf of the utility and then transfer
ownership of the asset to TANESCO, which may subse-
quently decide to convert the plant to run on natural gas.
Through this transaction, the capacity charge will fall to a
token amount, and following conversion to gas, the energy
charge will drop from $9 million–$12 million to $1 mil-
lion–$1.5 million a month. The PPA would also be termi-
nated and a new agreement drafted, one that would offer
customers discounted tariffs.
Project refinancing has limited application and must
be dealt with carefully during the project negotiation. As
one banker candidly indicated, “If project finance bankers
are expected to finance projects with the understanding
that periodically it will be necessary to have a restructur-
ing, the outcome of which is uncertain, the result will be
to eliminate the availability of nonrecourse financing.”
Given the already low level of availability of financing in
Africa, this should be avoided.
A host-country government’s willingness to share risks
over the life of an IPP may thus be pivotal to the long-term
sustainability of projects. That said, strategic management
does not occur in a vacuum. Often, the government is not
only an active counterparty but may even, as evidenced in
the refinancing of IPTL, initiate strategic management.
Other government-led initiatives include the government of
Tanzania’s buying down of Songas’ allowance for funds used
during construction. The myriad project-level factors are
summarized in table 21.4.
LESSONS LEARNED FROM IPPs IN
SUB-SAHARAN AFRICA
Despite numerous challenges, a number of Sub-Saharan
African countries have managed to attract and sustain pri-
vate investments in IPPs. More than 20 large IPPs have
taken root in eight countries, and a number of smaller, pri-
vate projects have also been developed. While some IPPs

Table 21.4 Elements Contributing to Successful IPP Investments

Element
Favorable equity partners
Favorable debt arrangements
Secure and adequate revenue stream
Credit enhancements and security arrangements
Positive technical performance
Strategic management and relationship building
Source: Authors.
Details
■ Local capital/partner contribution, where possible
■ Risk appetite for project
■ Experience with developing-country project risk
■ Involvement of a DFI partner and/or host-country government
■ Reasonable, fair return on equity
■ Firms with development origins
■ Competitively priced financing or involvement of DFIs
■ Local capital/markets to mitigate foreign exchange risk
■ Flexibility in terms and conditions (possible option to refinance)
■ Commercially sound metering, billing, and collections by the utility
■ Robust PPA (stipulates capacity and energy charges as well as dispatch, fuel
metering, interconnection, insurance, force majeure, transfer, termination, change of
law provisions, refinancing arrangements, dispute resolution)
■ Guarantees: sovereign guarantees, partial risk guarantees
■ Insurance: political risk insurance
■ Cash: escrow accounts, letters of credit, liquidity facilities
■ High level of technical performance (including availability)
■ Sponsors anticipate and mitigate potential conflicts (especially related to O&M and
budgeting)
■ Sponsors work to create good image in country through political relationships,
development funds, and effective communications and strategically manage their
contracts, particularly in the face of exogenous shocks and other stresses

386 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
have encountered some contract changes, nearly all have
survived and are countributing to social and economic
development.
At the country level, factors such as a favorable invest-
ment climate, clear policy and regulatory frameworks, and
local availability of cost-competitive fuels clearly help create
successful IPPs. Of growing importance are effective plan-
ning, procurement, and contracting policies and practices.
Kenya provides an example of how responsibilty for these
functions can be effectively allocated and institutionalized.
Although evidence is not conclusive, strategic manage-
ment on behalf of IPP sponsors and governments, as well as
strong technical performance, have been used to strengthen
projects. The role of firms with development origins, such
as Aldwych, Globeleq, and IPS, and development finance
institutions such as the IFC is increasingly important in the
successful development and operation of new IPPs in Sub-
Saharan Africa. Projects with participation of these firms
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
and DFIs were less likely to unravel, leading to one of two
conclusions: these projects may have been more balanced
from the outset and when an exogenous stress struck, or
they may have been better equipped to resist host-country
government pressure.
In Sub-Saharan African countries in which an imbalance
between development and investment outcomes is per-
ceived, there is evidence that IPP contracts are more
unlikely to unravel. The incidence of such unraveling, how-
ever, does not necessarily signal the end of a project’s oper-
ation. In all cases, efforts to close the initial gap between
investors and host-country governments’ perceptions and
treatment of risks must continue. Finally, the means of clos-
ing this gap may not be only, or mainly, through increasing
new protections such as PRGs or political risk insurance.
Rather, the solution may lie in systematic treatment of the
numerous elements contributing to success at the country
and project level defined in this chapter.
387
 ANNEX 21A PROJECT SPECIFICATIONS each of the regions are presented in alphabetical order.
Under each country, IPPs are then presented in chrono-
Detailed specifications are presented for each IPP discussed
logical order, based on the date when the IPP first came
in this chapter. Information is first grouped into regions
online.
with East Africa followed by West Africa. Countries within

East Africa
Kenya (1 of 5 projects)
Project Westmont
Size 46 MW
Cost US$65 million
$ per kW US$1,413
Fuel/technology Kerosene/gas condensate/gas turbine (barge mounted)
ICB None, but selective international tender conducted
Contract BOO, 7 years
Debt/equity —
DFI in equity and debt None
Local participation in equity and debt None
Equity partners (country of origin; % of each shareholder) Westmont (Malaysia, 100%) has sought to sell plant since 2004
Lenders —
Credit enhancements, security arrangements None
Project tender, COD 1996, 1997
Contract change Not a contract change per se, but firm failed to negotiate a second
contract after its 7 year contract ended in 2004 due to failure to
agree on tariffs
Fuel arrangement Originally Westmont to procure fuel and then pass through to utility,
however, following dispute with fuel supplier about taxes after the
first year of operation, utility took over procurement
Kenya (2 of 5 projects)
Project Iberafrica
Size 108.3 MW (44, 12, 52.3 brought on, respectively)
Cost US$35 million (only for first 56 MW)
$ per kW NA
Fuel/technology HFO/medium speed diesel engine
ICB None
Contract BOO, 7 years, 15 years, 25 years
Debt/Equity 72/28
DFI in equity and debt None
Local participation in equity and debt Yes (equity and debt)
Equity partners (country of origin; % of each shareholder) Union Fenosa (Spain, 80%), KPLC Pension Fund (Kenya, 20%) since 1997
Lenders Union Fenosa (US$12.7 million in direct loans and guaranteed
US$20 million); KPLC Staff Pension Fund (US$9.4 in direct loans and
guaranteed US$5 million through a local Kenyan bank).
Credit enhancements, security arrangements None
Project tender, COD 1996, 1997/1999, 2000/2008, 2009
Contract change Yes, Iberafrica reduced the capacity charge of its first PPA by 37% in
April 2002 and then to 59% of the original PPA in September 2003.a
Fuel arrangement Iberafrica buys fuel and passes cost through to KPLC based on the
units generated and specific consumption parameters agreed on in
the PPA
Kenya (3 of 5 projects)
Project OrPower 4
Size 13 MW + 35 MW
Cost US$105b million
$ per kW NA
Fuel/technology Geothermal
ICB Yes
(continued next page)

388 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
East Africa (continued)
Contract BOO, 20 years
Debt/equity NA
DFI in equity and debt DEG
Local participation in equity and debt None
Equity partners (country of origin; % of each shareholder) Ormat (100%) from 1998 to 2008.
Lenders —
Credit enhancements, security arrangements MIGA guarantee, a stand-by letter of credit, covering several months
billing (although only finalized at end-2006)
Project tender, COD 1996, 2000–2009
Contract change Yes, tariff for the second phased (35 MW) reduced
Fuel arrangement The only fuel arrangement per se is that OrPower 4 was granted a
Geothermal Resource License from the government, to which it pays
a royalty of sorts (of US$0.004/kWh or USc 0.4/kWh)
Kenya (4 of 5 projects)
Project Tsavo
Size 75 MW
Cost US$86 million
$ per kW US$1,133
Fuel/technology HFO/medium speed diesel engine
ICB Yes
Contract BOO, 20 years
Debt/equity 78/22
DFI in equity and debt Yes (equity and debt)
Local participation in equity and debt None
Equity partners (country of origin; % of each shareholder) Cinergy and IPS jointly owned 49. 9%, Cinergy sold to Duke Energy in
2005, CDC/Globeleq (U.K., 30%), Wartsila (Finland, 15%), IFC (5%)
retain remaining shares since 2000
Lenders IFC own account (US$16.5 million), IFC syndicated (US$23.5 million);
CDC own account (US$13 million), DEG own account
(€11 million), DEG syndicated (€2 million)
Credit enhancements, security arrangements Letter of comfort provided by government, and escrow account,
equivalent to 1 month capacity charge, and a stand-by letter of
credit, equivalent to 3 months billing
Project tender, COD 1995, 2001
Contract change None, but was pressured to lower tariff
Fuel arrangement Tsavo buys fuel and passes cost through to KPLC based on the units
generated and specific consumption parameters agreed on in the PPA
Kenya (5 of 5 projects)
Project Rabai
Size 90 MW
Cost US$155 million
$ per kW US$1,722
Fuel/technology HFO
ICB Yes
Contract BOOT, 20 years
Debt/equity 75% debt, (5% subordinated debt, 25% equity)
DFI in equity and debt Yes
Local participation in equity and debt No
Equity partners (country of origin; % of shareholder) Aldywch: 34%, BWSC (Danish, but owned by Mitsui of Japan): 25.5%,
FMO: 20%, IFU (Danish bilateral lender): 20%
Lenders DEG: 15%, FMO: 25%, EAIF: 25%, Proparco: 25%, EFP (European
Financing Partners): 10%
Credit enhancements, security arrangements Support letter from government of Kenya (covers political risk but falls
short of being an outright guarantee), and KPLC-issued letter of
credit equivalent to 5 months of capacity (debt service, fixed costs
and equity returns) payments and 2 months of fuel payments
Project tender, COD 2006, 2009
Contract change None
Fuel arrangement Fuel supply agreement with Kenol of Kenya
(continued next page)

CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 389
 East Africa (continued)
Tanzania (1 of 2 projects)
Project IPTL
Size 100 MW
Cost US$120 million
$ per kW US$1,200
Fuel/technology HFO/medium speed diesel engine
ICB None
Contract BOO, 20 years
Debt/equity 70/30 (albeit ratio disputed)
DFI in equity and debt None
Local participation in equity and debt Yes (in-kind equity participation)
Equity partners (country of origin; % of each shareholder) Mechmar (Malaysia, 70%), VIP (Tanzania, 30% in kind), both have sought
to sell shares
Lenders Two Malaysian-based banks (Bank Bumiputra Malaysia Berhad—now
Bank Bumiputra Commercial Bank—and SIME Bank) auctioned debt
to Standard Chartered Bank
Credit enhancements, security arrangements Sovereign guarantee, liquidity facility equivalent to 4 months capacity
charge (but not yet established)
Project tender, COD 1997, 2002
Contract change Yes, postarbitration, monthly capacity charges lowered from US$3.6
million to US$2.6 million
Fuel arrangement IPTL imports fuel, which is a pass-through to the utility
Tanzania (2 of 2 projects)
Project Songas
Size 180 MWc
Cost US$316 milliond
$ per kW US$2,313 (for first 115 MW) and US$769 (for 65 MW expansion)
Fuel/technology Natural gas/open cycle
ICB Yes
Contract BOO, 20 years
Debt/equity 70/30 for 115 MW
100% equity financed for 65 MW expansion, which has since been
refinanced
DFI in equity and debt Yes
Local participation in equity and debt Yes (TANESCO, TDFL)
Equity partners (country of origin; % of each shareholder) TransCanada sold majority shares to AES (USA) in 1999 and AES sold
majority shares to Globeleq (UK) in 2003.e All preferred equity
shares were converted into “Loan Notes” in June 2009. Only
common shares remain.
Lenders World Bank (US$136 million), EIB (US$55 million), Sida (US$15 million)
Credit enhancements, security arrangements Escrow account: for first 115 MW, with the government matching every
US$1 spent by the project company; liquidity facility equivalent to
4 months capacity charge for the first 3 years, declining to 2 months
starting in year 4 through the remaining years of the contract
Project tender, COD 1994, 2004
Contract change The changes to Songas’ PPA include: US$103 million AFUDC buy-
down; the use of escrow funds to buy down the AFUDC; freezing
the liquidity facility, which was meant to be replenished and accessed
only in the case of partial or nonpayment by TANESCO; and
preferred equity conversion to “loan notes”
Fuel arrangement Songo Songo gas provided to project company at a rate of US$0.55/
MMBtu for turbines I–V and at US$2.17 MMBtu for turbine VI
Uganda (1 of 2)
Project Namanve Power Plant/Jacobsen Uganda Power Plant Limited
Size 50 MW
Cost US$74 million at inception
$ per kW US$1,488
Fuel/technology HFO
ICB Yes
(continued next page)

390 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
East Africa (continued)
Contract BOOT, 6 year
Debt/equity —
DFI in equity and debt Yes (debt)
Local participation in equity and debt Yes (debt)
Equity partners (country of origin; % of each shareholder) Jacobsen Elektro (Norway, 100%)
Lenders Loan from Standard Chartered Bank UK, partly guaranteed by GIEK
(the Norwegian guarantee agency), commercial loan from Stanbic
Bank (Uganda) Ltd. (backstopped by Standard Chartered Bank UK);
subordinate loan from parent company Jacobsen Elektro AS;
NORAD grant
Credit enhancements, security arrangements GIEK guarantee backed up by government of Uganda
Project tender, COD COD 2008
Contract change None
Fuel arrangement Openly bid for contract; not part of license (plant availability should
be >90%)
Uganda (2 of 2)
Project Bujagali
Size 250 MW
Cost US$860 million
$ per kW US$3,440
Fuel/technology Hydro
ICB Yes
Contract BOT, 30 years
Debt/equity 78/22
DFI in equity and debt Debt
Local participation in equity and debt Yes, government of Uganda
Equity partners (country of origin; & % of each shareholder) Sithe Global (58%), IPS/Aga Khan (32% but 50.1% voting), government
of Uganda (10%)
Lenders African Development Bank, EIB, PROPARCO, AFD, DEG, KfW, FMO, and
IFC as DFIs, NedBank, Absa Capital, Standard Chartered Bank, and
Fortis as commercial banks covered by the PRG
Credit enhancements, security arrangements Government guarantee, MIGA, PRG/IDA
Project tender, COD 2005, expected 2012
Contract change None
Fuel arrangement —

West Africa
Côte d’Ivoire (1 of 2)
Project Compagnie Ivoirienne de Production d’Electricité (CIPREL)
Size 210 MW
Cost US$105.6 millionf
$ per kW US$503
Fuel/technology Natural gas/open cycle
ICB None
Contract BOOT, 19 years
Debt/equity —
DFI in equity and debt Yes (equity and debt)
Local participation in equity and debt None
Equity partners (country of origin; % of each shareholder) SAUR International, with 88% joint venture (JV) between French SAUR
Group owned by Bouygues, 65% and EDF, 35%) BOAD, PROPARCO,
and IFC holding the remaining 12%; in 2005 all shares sold to
Bouygues (France, 98%), except BOAD (2%)
Lenders World Bank
Credit enhancements, security arrangements None
Project tender, COD 1993, 1995
Contract change None
Fuel arrangement Government procures fuel
(continued next page)

CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 391
 West Africa (continued)
Côte d’Ivoire (2 of 2)
Project Azito
Size 288 MWg
Cost US$225 million
$ per kW US$781
Fuel/technology Natural gas/open cycle
ICB Yes
Contract BOOT, 24 years
Debt/equity 70/30
DFI in equity and debt Yes (debt)
Local participation in equity and debt None
Equity partners (country of origin; % of each shareholder) Cinergy (JV between Swiss ABB, 50% and French EDF, 50%)
Holds 65.7% of shares, CDC/Globeleq (11%), and IPS (23%)
Lenders IFC, CDC, FMO, DEG, AfDB, Societe Generale, and other European
commercial banks
Credit enhancements World Bank partial risk guarantee
Sovereign guarantee
Security arrangements Escrow account equivalent to 1 month capacity charge
Project tender, COD 1996, 2000
Contract change None
Fuel arrangement Government procures fuel
Ghana (1 of 3)
Project Takoradi II
Size 220 MWh
Cost US$110 million
$ per kW US$500
Fuel/technology Light crude oil/single cycle
ICB None
Contract BOOT, 25 years
Debt/Equity Financed exclusively with balance sheet of sponsors
DFI in equity and debt No
Local participation in equity and debt Yes (equity)
Equity partners (country of origin & % of each shareholder) CMS (U.S., 90%), VRA (Ghana, 10%), CMS sold shares to TAQA
(UAE, 90%) in 2007
Lenders Financed exclusively with balance sheet of sponsors
Credit enhancements Sovereign guarantee
Security arrangements US$3 million letter of credit provided by government
Project tender, COD 1998, 2000
Contract change Failure to develop second phase (110 MW), investors cite the lack of
government guarantees (granted in the first phase of Takoradi II),
meanwhile government has indicated that the EPC costs are too high
and further development remains at standstill
Fuel arrangement Government procures fuel
Ghana (2 of 3)
Project Sunon Asogli Power Plant
Size 200 MW
Cost —
$ per kW —
Fuel/technology Combustion engine
ICB No
Contract BOO, 20 years
Debt/equity —
DFI in equity and debt None
Local participation in equity and debt Local strategic investor, Togbe Afede XIV
Equity partners (country of origin; % of each shareholder) Shenzhen (China, 100%)
Lenders —
Credit enhancements, security arrangements —
Project tender, COD —
(continued next page)

392 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
West Africa (continued)
Contract change Related to fuel
Fuel arrangement Inability to reach conclusion about fuel agreement, which has prevented
plant from coming online
Ghana (3 of 3)
Project Bui hydro
Size 400 MW
Cost US$622 million
$ per kW US$1,555/kW
Fuel/Technology Hydro
ICB No
Contract —
Debt/Equity —
DFI in equity and debt None
Local participation in equity and debt —
Equity partners (country of origin & % of each shareholder) Sinohydro (China, 100%)
Lenders —
Credit enhancements and security arrangements —
Project tender, COD 2005, expected online in 2012
Contract change None
Fuel arrangement Hydro
Nigeria (1 of 4)
Project AES Barge Limited
Size 270 MW
Cost US$240 million
$ per kW US$888
Fuel/technology Natural gas/open cycle (barge mounted)
ICB None
Contract BOO, 13 years
Debt/equity —
DFI in equity and debt None
Local participation in equity and debt Yes (equity)
Equity partners (country of origin; % of each shareholder) Enron (U.S., 100%) sold to AES (95%) and YFP (Nigeria, 5%) in 2000
Lenders —
Credit enhancements, security arrangements OPIC political risk insurance
Sovereign guarantee, US$60 million letter of credit from Ministry
of Finance
Project tender, COD 1999, 2001
Contract change Yes, initial plant size increased from 90 MW to 270 MW (9 units of 30
MW each) and change in the fuel from liquid fuel to natural gas, both
of which had the effect of reducing the capacity charge, current
contract changes under discussion involve among other things the
availability of deficiency payment, meanwhile tax exemption certificate
has been withheld by government for the duration of the project
Fuel arrangement Utility arranges fuel
Nigeria (2 of 4)
Project Okpai
Size 450 MW
Cost US$462i
$ per kW —
Fuel/technology Natural gas/combined cycle
ICB None
Contract BOO, 20 years
Debt/equity 100% equity financed
DFI in equity and debt None
Local participation in equity and debt Yes (equity and debt)
Equity partners (country of origin; % of each shareholder) Nigerian National Petroleum Corporation (Nigeria, 60%), Nigerian
Agip Oil Company (Italy, 20%), and Phillips Oil Company (U.S., 20%)
maintained equity since 2001
(continued next page)

CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 393
 West Africa (continued)
Lenders Provided by equity partners
Credit enhancements, security arrangements PPA backed by Nigerian Petroleum Development Company’s oil
revenues
Project tender, COD 2001, 2005
Contract change Ongoing negotiations related to investment costs which rose by
US$150 million, to US$462 million; although plant is producing
power, due to the dispute, full payment is not being made by utility
Fuel arrangement Project company provides fuel
Nigeria (3 of 4)
Project Afam IV
Size 630 MW
Cost —
$ per kW —
Fuel/technology CCGT
ICB No
Contract BOO, 20 years
Debt/equity 100% equity
DFI in equity and debt None
Local participation in equity and debt Yes, NNPC
Equity partners (country of origin; % of each shareholder) NNPC (Nigeria, 55%), Shell (U.K./Netherlands, 30%), Elf (Total)
(France, 10%), Agip (Italy, 5%)
Lenders
Credit enhancements, security arrangements PPA backed by Nigerian Petroleum Development Company’s
oil revenues
Project tender, COD 2000, 2007
Contract change No
Fuel arrangement Project company provides fuel
Nigeria (4 of 4)
Project Aba Integrated Power Project
Size 140 MW
Cost US$385
$ per kW US$2,750/kW (includes distribution investment)
Fuel/t\Technology Gas fired open-cycle power
ICB No
Contract BOO
Debt/equity
DFI in equity and debt No
Local participation in equity and debt Yes (equity)
Equity partners (country of origin; % of each shareholder) Geometric Power Limited (Nigeria)
Lenders Diamond Bank, Stanbic IBTC
Credit enhancements, security arrangements
Project tender, COD 2005, anticipated COD end-2010
Contract change No
Fuel arrangement
Senegal (1 of 2)
Project GTi Dakar
Size 52 MW
Cost US$65 million
$ per kW US$1,250/kW
Fuel/technology Complete cycle/diesel/natfa
ICB Yes
Contract BOOT, 15 years
Debt/equity 75/25
DFI in equity and debt Yes (both)
Local participation in equity and debt None
Equity partners (country of origin, % of each shareholder) GE Capital Structured Finance Group (U.S.), IFC, Edison (Italy)
Lenders IFC and Credit Commercial de France
(continued next page)

394 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
West Africa (continued)
Credit enhancements, security arrangements Government guarantee, credit insurance through a guarantee program
of SACE, the Italian export credit agency, and a partial interest
subsidy through the Mediocredito Central Subsidy Department,
escrow account
Project tender, COD 1996, 1999/2000
Contract change None
Fuel arrangement
Senegal (2 of 2)
Project Kounoune I
Size 68 MW
Cost US$110 million
$ per kW US$1617/kW
Fuel/technology HFO diesel
ICB Yes
Contract BOO, 15 years
Debt/equity
DFI in equity and debt Yes (debt)
Local participation in equity and debt Yes (debt)
Equity partners (country of origin; % of each shareholder) Mitsubishi (Japan), Matelec S.A.L (Lebanon)
Lenders IFC, Proparco, AfDB, BOAD, and CBAO
Credit enhancements, security arrangements Government guarantee, a PRG although never signed by the
government, a letter of credit from Senelec
Project tender, COD 2003, 2008
Contract change None
Fuel arrangement During the project negotiations, the structures of the FSA and PPA
were changed to turn the PPA into a tolling agreement.
Togo (1 of 1)
Project Extension of Central Thermique de Lome (CTL)
Size 100 MW
Cost US$196 million (including 18 km transmission lines rehabilitation and
soil decontamination investment)
$ per kW US$1,960
Fuel/technology Triple fuel
ICB No
Contract BOOT, 25 years
Debt/equity 75/25
DFI in equity and debt Yes (debt)
Local participation in equity and debt None (but contract stipulates up to 25% of equity must be sold to
local/Togolese)
Equity partners (country of origin; % of each shareholder) Contour Global (U.S., 80%), IFC (20%)
Lenders OPIC
Credit enhancements, security arrangements Government guarantee, OPIC guarantee, escrow account for 1 month
of full operation, letter of credit of 2 months of full operation, (which
become one month after 2 years of no incident of payment, and
ceases to exist after 4 years of no incident)
Project tender, COD Due online in July 2010
Contract change No
Fuel arrangement The buyer (utility provides fuel)

Source: Authors.
Notes: — = Not available. BOT = build, own, transfer.
a. Furthermore, although not a contract change per se, the value of the capacity charge for Iberafrica’s second PPA was 50 percent that of the first PPA.
b. US$105 includes only the loan portion for the 35 MW part of the plant (Ormat, personal communication, 2010).
c. There was considerable evolution in terms of the planned capacity for the plant, from 60 MW to the current 180 MW.
d. Songas project costs include refurbishment of gas wells, a new gas processing facility, pipeline construction and fuel conversion of the existing
power station (Ubungo), in total amounting to US$266 million, and an additional US$50 million for expansion in terms of two additional turbines
(total 65 MW) and related infrastructure. The expansion was financed entirely by equity. A rough estimate for the electricity generation component
would be 40 percent of project costs or US$130 million, based on US$35 million for refurbishment and fuel conversion of existing turbines, US$45
million assumed loans on existing turbines, and US$50 million for expansion.
(continued next page)

CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 395
 West Africa (continued)
e. Due to complexity, turnover is detailed in this footnote: Ocelot (Canada), TransCanada (Canada), Tanzania Petroleum Development Corpora-
tion, TPDC (Tanzania), TANESCO (Tanzania), Tanzania Development Finance Company, TDFL (Tanzania, sponsored by EIB), IFC (multilateral),
DEG (German), CDC (UK) were shareholders by 1996, with TransCanada the majority shareholder; IFC and DEG sold shares to CDC in 1997/8;
TransCanada sold shares to AES (USA) in 1999; Ocelot/PanOcean sold shares to AES in 2001; AES sold majority shares to Globeleq (UK) and FMO
(Holland) in 2003. After the AES sale, equity shares and associated financial commitments (expressed in US$ million) in Songas were as follows:
Globeleq: US$33.8 (56%); FMO: US$14.6 (24%); TDFL: US$4 (7%); CDC: US$3.6 (6%); TPDC: US$3 (5%) and TANESCO: US$1 (2%). This does
not reflect the additional US$50 million that Globeleq committed for the expansion.
f. Investment cost €87.8m or 57.6 billion CFA, with the average 1994 conversion of US$ to CFA, 545,100.
g. The initial project concept included specifications to raise capacity to 420 MW.
h. The initial project concept included specifications to add a second phase of 110 MW and convert to combined cycle, however, lack of funding
has limited the completion of this phase.
i. Project costs include the gas infrastructure.

NOTES continuous power producers, privately owned by the sugar

1. The standard model for power sector reform has been
roughly defined as a series of steps that move vertically
integrated utilities toward competition, generally includ-
ing the following activities: corporatization, commercial-
ization, passage of the requisite legislation, establishment
of an independent regulator, introduction of IPPs,
restructuring/unbundling, divestiture of generation and
distribution assets, and introduction of competition
(Bacon 1999; Adamantiades and others 1995; Besant-Jones
2006; Williams and Ghanadan 2006). Although this model,
based largely on the early power sector reforms carried out
in Chile, Norway, and the United Kingdom, came to repre-
sent a standard, it is arguable that not all steps were relevant
to conditions on the ground in most developing countries
(Gratwick and Eberhard 2008b).
2. Exceptions are Côte d’Ivoire and Tanzania, where IPPs
are contributing more than 50 percent to overall electricity
production. Togo’s first IPP, Centrale thermique de Lomé,
will also supply the majority of electricity for that country.
3. Two different categories of DFIs should be distinguished,
namely, those that lend on commercial terms and largely to
private companies (for example, FMO, PROPARCO, DEG,
and IFC), and the multilateral development banks (such as
the World Bank and the African Development Bank) that
lend on concessionary terms and primarily to public sector
projects. It is the latter that refocused on infrastructure
(Rudo 2010a).
4. Although Mauritius has four IPPs (which, at approxi-
mately 200 MW combined, account for about 37 percent of
installed capacity and a little less than 25 percent of pro-
duction, as of end-2005), the country has not been
included in this sample. The IPPs, which are all cogenera-
tion plants, provide power and steam to the country’s sugar
mills throughout the crop season, reducing their contribu-
tion to the state-owned utility by about 30 percent. During
this time, the shortfall in production is made up by seven
mills. With installed capacity of 40 MW, roughly equal to
the IPP shortfall, these continuous producers also have
long-term take or pay contracts with the state (Bergesen
2007; World Bank 2007a).
5. Centrale thermique de Lomé is a preexisting 90 MW
facility for which a rehabilitate-own-transfer contract was
provided to Electro Togo in 2001. In 2006, however, when
a power crisis intensified and the rehabilitation had not
yet been completed, the government of Togo opted to ter-
minate Electro Togo’s contract and initiate a BOOT
directly with Contour Global, which had previously been
involved in discussions with Electro Togo (Ministry of
Energy of Togo 2010).
6. A portion of the shares of the two Kenyan utilities, KPLC
and Kengen, however, are listed on the Nairobi Stock
Exchange. Côte d’Ivoire’s state utility has been under a pri-
vate concession contract since 1990—an arrangement that is
expected to continue until 2020. TANESCO in Tanzania and
KPLC in Kenya also have had private management contracts.
In Senegal, there was an attempt (albeit aborted) to privatize
Senelec in 1999 through a concession with Hydro-Québec
and Elyo (the concession was annulled by the president in
2001). Uganda unbundled its national utility into separate
generation, transmission, and distribution companies and
entered into long-term private concession in generation
(Eskom) and distribution (Umeme-Globeleq). Finally,
Nigeria has expressed an intention to unbundle and priva-
tize but has not yet fully realized this ambition.
7. Alternatives to strictly independent regulation that may
provide a better match to a country’s regulatory commit-
ment and institutional and human resource capacity (for
example, regulatory contracts, outsourcing of regulatory
functions, expert panels, and regional regulators) are
increasingly being considered (Eberhard 2007).
8. Domestic gas reserves were used for IPPs in Côte
d’Ivoire. Unlike for the other countries, however, this use

396 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
did not represent the establishment of a new gas infra-
structure. An attempt, led by the government, was also
made to exploit stranded gas reserves in Ghana’s Osagyefo
Barge project.
9. Stakeholders in Nigeria have seen costs escalate, causing
the utility to withhold payments. In Tanzania costs have also
escalated, but for reasons unrelated to the project itself.
10. At the time of writing, Kenya had only 60 MW still in
service, and that was expected to be retired by November
2010.
11. In terms of international norms, however, it should be
noted that Tanzania’s cost of unserved energy (CUE) is low.
South Africa’s CUE is approximately US$10/kWh, which is
in line with the CUE in many industrialized countries
(Global Energy Decisions 2007).
12. On the other hand, given the high cost of mounting a
bid, having an excessive amount of bidders would also not
be desirable as some will not bid if they perceive their
chances to be low (Rudo 2010a).
13. Togo’s next IPP, a 24 MW wind farm, is being procured
through an ICB (Ministry of Energy of Togo 2010).
14. Tanzania used the net-back calculation method to estab-
lish the gas price for the first 142 MW owned by Songas.
Although it is viewed to be low (“US$0.55/Mcf ”), most of
the costs associated with the gas infrastructure are paid sep-
arately as part of the capacity charge. In typical projects,
these would have been part of the gas price (i.e.,
US$4.20/Mcf). The same gas price is not offered to IPTL.
Under the IPTL PPA, fuel is a pass-through. TANESCO has
negotiated with PanAfrican Energy and TPDC to book 100
MMcfd at negotiated wellhead prices (in favor of the addi-
tional 40 MW owned by Songas, 145 MW owned by
TANESCO, and the next 100 MW plant, which may be IPTL
or any other plant) starting at US$1.98/Mcf, escalated at 2.2
percent annually for the first five years from July 2007, then
to US$3.5/Mcf thereafter, escalated at USCPI. In summary,
the gas prices are discounted market prices of an alternative
fuel (EWURA 2010a).
15. Projects such as Osagyefo Barge in Ghana and Ibom
and Omoku in Nigeria are sometimes loosely termed
IPPs. In Ghana, private participation was expected (but
not achieved), and in Nigeria, projects, although inde-
pendent of the national utility, have been led entirely by
the Rivers and Akwa Ibom state governments, respec-
tively. All of these projects are excluded from the analysis
presented here.
16. At the time of writing, the level of equity (including
local equity) is under dispute in Tanzanian court.
17. The emergence of two Chinese firms, Shenzhen and
Sinohydro, in Sunon Asogli and Bui hydro, respectively, in
Ghana could represent a new element and be a significant
source of investment for other African IPPs. Foster and
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
Briceno-Garmendia (2010) report on the general phenom-
enon of non-OECD funders, including the emergence of
increasingly significant funding flows from China to Sub-
Saharan Africa.
18. In 2009, 10 years after financial close, DEG arranged a
$105 million loan (Ormat 2009).
19. In 2005, it was found that the then managing director of
KPLC, Samuel Gichuru, received $2 million from Westmont.
20. In addition, a PRG for Kounoune I was approved by the
World Bank but was ultimately not signed by the govern-
ment of Senegal, because the lender that the PRG would
have covered was willing to fund the project without the
PRG (IFC 2010a).
21. World Bank 2010. For IPPs in Central Asia, however,
and for cross-border projects such as those in Laos and
Thailand (Nam Theum II), a PRG or other “strong” credit
enhancements would be employed. In the middle-income
Latin American countries cited above, privatization trajec-
tories meant there were few IPPs, and in those few cases (for
example, Colombia), bilateral and multilateral institutions
helped back investments (Rudo 2010a).
22. The fact that Iberafrica was not project-financed meant
that the company had greater flexibility to change the pay-
ment streams.
REFERENCES
Adamantiades, A. G., et al. 1995. “Power Sector Reform in
Developing Countries and the Role of the World Bank.”
Paper prepared for 16th Congress of the World Energy
Council, Tokyo. World Bank, Industry and Energy
Department, Washington, DC.
Africa Electra. 2005. “The Role of Local Banks in Financing
Power Projects.” Africa Electra 30.
Aldwych International. 2010a. Personal communication,
May 5–6.
———. 2010b. Personal communication, June 9.
APEC (Asia-Pacific Economic Cooperation) Energy Working
Group. 1997. “Manual of Best Practice Principles for Inde-
pendent Power Producers.” APEC Secretariat, Canberra.
Bacon, R. 1999. “A Scorecard for Energy Reform in Devel-
oping Countries.” World Bank, Finance Private Sector
and Infrastructure Network, Washington, DC.
Bergesen, C. 2007. Mauritius. Columbus, OH: McGraw Hill.
Besant-Jones, J. E. 2006. “Reforming Power Markets in
Developing Countries: What Have We Learned?” Energy
and Mining Sector Board Discussion Paper 19, World
Bank, Washington, DC.
Deloitte Touche Tohmatsu Emerging Markets Ltd. and
Advanced Engineering Associates International. 2003.
“Energy Program South Asia Generation Pricing Study.”
397
 South East Asia Regional Initiative for Energy, USAID,
Washington, DC.
Eberhard, A. 2007. “The Independence and Accountability
of Africa’s Infrastructure Regulators: Reassessing Regu-
latory Design and Performance.” Paper presented at
AFUR Annual Conference, Livingstone, Zambia, April
25–26.
Eberhard, A., and I. Malgas. Forthcoming. “Hybrid Power
Markets in Africa: Generation Planning, Procurement
and Contracting Challenges.” Energy Policy.
Eberhard, A., and M. Shkaratan. 2010. “Africa’s Power Infra-
structure: Improving Investment, Connectivity, Reliabil-
ity and Efficiencies.” Africa Infrastructure Country Diag-
nostic, World Bank, Washington, DC.
EDF (Électricité de France). 2005. Personal communication,
January 11.
ERC (Energy Regulatory Commission of Kenya). 2010a.
Personal communication, May 18. Nairobi, Kenya.
———. 2010b. Personal communication, May 26. Nairobi,
Kenya.
EWURA (Energy and Water Utilities Regulatory Author-
ity). 2010. Personal communication, May 18. Dar es
Salaam Tanzania.
Foster, V., and C. Briceño-Garmendia, eds. 2010. Africa’s
Infrastructure: A Time for Transformation. Washing-
ton, DC: Agence Française de Développement and
World Bank.
Global Energy Decisions. 2007. “Electrical Resource Needs
Analysis: Adequate Reserve Margin for Development of
Third National Integrated Resource Plan for South
Africa.” Pretoria.
Globeleq. 2006. “Enabling Electricity Growth in Africa:
Some Experiences.” In Financing Electricity for Growth in
Africa. Tunis: Infrastructure Consortium for Africa.
———. 2010a. Personal communication, May 5.
———. 2010b. Personal communication, June 1.
Gratwick, K. N., and A. Eberhard. 2008a. “An Analysis of
Independent Power Projects in Africa: Understanding
Development and Investment Outcomes.” Development
Policy Review 26 (3): 309–38.
———. 2008b. “Demise of the Standard Model for Power
Sector Reform and the Emergence of Hybrid Power
Markets.” Energy Policy 36: 3948–60.
IFC (International Finance Corporation). 2005. Personal
communication, January 24.
———. 2010a. Personal communication, May 21. Washington,
DC.
———. 2010b. Personal communication, June 12. Washington,
DC.
398 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
IPS (Industrial Promotion Services). 2010. Personal com-
munication, May 24. Nairobi, Kenya.
Hoskote, M. 1995. “Independent Power Projects (IPPs): An
Overview.” Energy Note 2. In Energy Themes, ed. R. English.
Washington, DC: World Bank.
International Energy Agency. 2006. World Energy Outlook
2006. Paris: OECD/IEA.
Malgas, I., K. N. Gratwick, and A. Eberhard. 2007a. “Moroc-
can Independent Power Producers: African Pioneers.”
Journal of North African Studies 13 (1): 15–36.
———. 2007b. “Two of a Kind: Lessons from Tunisian
Independent Power Projects.” Journal of North African
Studies 12 (4): 395–415.
Matsukawa, T., R. Sheppard, and J. Wright. 2003. “Foreign
Exchange Risk Mitigation for Power and Water Projects
in Developing Countries.” Energy and Mining Sector
Board Discussion Paper 9, World Bank, Washington, DC.
Ministry of Energy of Togo. 2010. Personal communication,
June 10–11. Lome, Togo.
Nigeria Electricity Regulatory Commission. 2010. Per-
sonal communication, May 21. Abuja, Nigeria.
Ormat. 2009. “Ormat Technologies Announces Closing of
$105 Million Long-Term Senior Debt Financing for the
Olkaria III Geothermal Power Project in Kenya.” Press
release, March 6, Reno, Nevada.
———. 2010. Personal communication, May 17.
Phadke, A. 2007. “Mark-ups in the State Capital Cost of
Private Power Projects in Developing Countries.”
Energy and Resources Group, University of California,
Berkeley.
PURC (Public Utilities Regulatory Commission of Ghana).
2010. Personal communication, May 19–21. Accra, Ghana.
Regulatory Commision of Senegal. 2010. Personal commu-
nication, June 11. Dakar, Senegal.
Rudo, D. 2010a. Personal communication, May 14. Washington,
DC.
———. 2010b. Personal communication, June 10. Washington,
DC.
Sinclair, S. 2007. “The World Bank Group Risk Mitigation
Instruments.” Paper presented at Private Sector Partici-
pation in Power Conference, Dead Sea, Jordan.
UNEP (United Nations Environment Programme) and UN
Economic Commission for Africa. 2006. Making Africa’s
Power Sector Sustainable: An Analysis of Power Sector
Reforms in Africa. New York: United Nations.
Williams, J. H., and R. Ghanadan. 2006. “Electricity Reform
in Developing and Transition Countries: A Reappraisal.”
Energy 31: 815–44.
Woodhouse, E. J. 2005. “The Experience with Independent
Power Projects in Developing Countries: Interim
 Report.” Program on Energy and Sustainable Develop-
ment, Stanford University, Stanford, CA.
Woolf, F., and J. Halpern. 2001. “Integrating Independent
Power Producers into Emerging Wholesale Power Mar-
kets.” Policy Research Working Paper 2703, World Bank,
Washington, DC.
World Bank. 1993. The World Bank’s Role in the Electric
Power Sector: Policies for Effective Institutional, Regulatory,
and Financial Reform. Washington, DC: World Bank.
———. 1997. “Morocco’s Jorf Lasfar Power Station.” In Proj-
ect Finance and Guarantees. World Bank, Washington, DC.
CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS
———. 1999. “Sub-Saharan Africa Benefits from the first
IDA Guarantee for Azito.” In Project Finance and Guar-
antees. Washington, DC: World Bank.
———. 2007. Private Participation in Infrastructure Data-
base. http://ppi.worldbank.org.
———. 2010. Personal communication, May 17.
World Bank and USAID. 1994. “Submission and Evaluation
of Proposals for Private Power Generation Projects in
Developing Countries.” IEN Occasional Paper 2. World
Bank, Washington, DC.
399
PA R T V I

Improving Human Development Outcomes

with Innovative Policies
 CHAPTER 22
Innovative Financing for Health in Rwanda:
A Report of Successful Reforms
C. Sekabaraga, A. Soucat, F. Diop, and G. Martin
wanda has been moving very rapidly to expand
R
health service delivery. It has dramatically acceler-
ated the trend of progress on health indicators,
putting it back on track to reach the health Millennium
Development Goals (MDGs). Three interrelated innovative
reforms contributed to improvement: community-based
health insurance, performance-based financing within a
broader framework of reform of management of human
resources for health, and fiscal decentralization. This chap-
ter examines these reforms, showing how they worked
together to improve health outcomes in Rwanda.
Health indicators in Rwanda improved dramatically in
recent years, and projections for the coming years are pos-
itive. After an initial surge following the genocide, under-
five mortality (the probability of death per 1,000 live
births) significantly decreased, falling from 196 in 2000 to
152 in 2005 and 103 in 2007 (figure 22.1). The infant mor-
tality rate decreased from 107 per 1,000 live births in 2000
to 86 in 2005 and 62 in 2007. Improvements were particu-
larly significant among the poor, with the under-five
mortality rate among the poorest quintile declining by
50 deaths per 1,000 between 2005 and 2008 compared with
38 deaths per 1,000 live births among the richest quintile.
The annual rate of decrease achieved between 2005 and
2008 was 12.2 percent—far greater than the 9.7 percent
needed to meet the MDG 4 target of 50 in 2015. Rwanda is
thus back on track to reach the health MDGs.
Since 2000 the maternal mortality ratio has declined at
an annual rate of 12.1 percent to reach 383 per 100,000 in
2008, ranking it among the best-performing countries in
the world (figure 22.2). This rate of decline far exceeds the
5 percent rate needed to meet the MDG target of reducing
the maternal mortality ratio by three-quarters between 1990
and 2015.
These achievements have been the result of innovative
strategies to address some of the key challenges affecting
maternal mortality. The share of women delivering their
babies in health facilities has steadily increased, rising from
28 percent of pregnant women in 2000 to 45 percent in 2008.
Many challenges remain, but preliminary data from the
Ministry of Health for 2010 suggest that this figure has risen
to two-thirds of all pregnant women; the finding should be
validated by the 2010 Demographic and Health Survey.
These successes can be attributed largely to an increase in
the use of essential health interventions, particularly high-
impact interventions that are critical in reducing disease
burden in developing countries, including immunization,
assisted deliveries, family planning, and the use of insecticide-
treated bed nets to prevent malaria. Rwanda has maintained
a very high and equitable coverage of vaccination against
avoidable childhood diseases since 2000: immunization rates,
at 95 percent in 2008, are among the highest in Sub-Saharan
Africa. Major progress has also been made in extending the
coverage of vitamin A supplementation among children and
women, through a mass campaign and integration into rou-
tine health facility services. Treatment of acute respiratory
infections of children has also increased, including among
the poor.
403
 Figure 22.1 Under-five Mortality Rate in Rwanda, 1990–2015

250

200 196

Deaths per 1,000

151 152
150 MDG

103
100
50
50

0
90

92

94

96

98

00

02

04

06

08

10

12

14
19

19

19

19

19

20

20

20

20

20

20

20

20
Sources: Measure DHS and ICF Macro 1995, 2000, 2005, 2008.

Figure 22.2 Maternal Mortality Ratio and Facility-based Deliveries in Rwanda, 2000–15

100
Maternal mortality ratio (per 100,000)

1,200 1,071
90

Facility-based deliveries (%)

1,000 80
750 70
800 66
60
600 50
45
40
400 383
26 28 268 30
200 20
10
0 0
00

01
02
03

04
05
06
07
08
09
10
11
12

13
14
15
20

20
20
20

20
20
20
20
20
20
20
20
20

20
20
20

Maternal mortality ratio Facility-based deliveries

Sources: Measure DHS and ICF Macro 2005, 2008.

Improvements in the use of women’s health services are HIV/AIDS epidemic. Malaria incidence and mortality have
also evident, with significant increases in the proportion of declined dramatically, largely as the result of increased use
assisted birth deliveries and the number of emergency of insecticide-treated nets, which among children under
obstetrical cases referred. The use of modern contraceptives five rose from 11 percent in 2005 to 56 percent in 2007. As
increased from 3 percent in 2000 to 27 percent in 2007, one a result, malaria-specific mortality was cut in half. Rwanda
of the fastest increases ever observed. The proportion of has thus moved from being a country where malaria was
women having at least one antenatal consultation rose from endemic to one focusing on eliminating malaria as a pub-
58 percent in 1995–2000 to 96 percent in 2000–07. The pro- lic health problem. The HIV/AIDS epidemic has been con-
portion of assisted deliveries increased from 39 percent in tained, with 3 percent of the population affected and more
2005 to 52 percent in 2007. than 60 percent of patients needing treatment receiving
Major progress has also been made in controlling com- highly active antiretroviral therapy. Knowledge of
municable diseases—including malaria, a prime cause of HIV/AIDS is better in Rwanda than in any other Sub-
morbidity and mortality in Rwanda—and containing the Saharan African country: nationwide 54 percent of women

404 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
and 58 percent of men had comprehensive knowledge of
HIV/AIDS in 2005.1
This progress occurred in a context in which annual
total per capita health expenditures doubled, from $17
to $34 between 2003 and 2006. In 2006 total health
expenditures reached 10.7 percent of gross domestic
product (GDP) (one of the highest levels of health expen-
ditures observed in low-income countries), up from
3 percent in 2002.
The overall share of public expenditures allocated to
health has remained stable, with most of the increase com-
ing from donors. Private expenditures were about $9.40 per
capita, and domestic public expenditure about $6.30 per
capita, with donors contributing $17.70 per capita, one of
the highest levels of donor dependency in Sub-Saharan
Africa. Much of this donor funding is earmarked funding
for HIV/AIDS.
Despite the increase in funding, resources are not suf-
ficient to meet the country’s health care needs. Rwanda
has therefore pioneered profound reforms, including an
innovative health system and a financing model grounded
in grassroots initiatives and institutions. Three prominent
reforms were adopted to boost both the demand for and
the supply of health services: health microinsurance
(mutuelles), performance-based financing, and fiscal
decentralization. Those reforms have transformed the fis-
cal space landscape. Revenues generated by health facili-
ties have increased significantly as a result of increased
use of health services and health insurance coverage, and
an increasing share of domestically generated revenues is
captured by health centers, strengthening front-line
providers.
In 2007 the Rwandan government adopted its second
Poverty Reduction Strategy Paper (the Economic Develop-
ment and Poverty Reduction Strategy). Its goals for the
health sector are to maximize preventive health measures
and build the capacity to provide high-quality and accessi-
ble health care services to the entire population in order to
reduce malnutrition, infant and child mortality, and fertility
and to control communicable diseases. The strategy also
supports strengthening institutional capacity, increasing the
quantity and quality of human resources, ensuring that
health care is accessible to the entire population, increasing
geographical accessibility, increasing the availability and
affordability of drugs, improving the quality of services in
the control of diseases, and encouraging the demand for
such services. It also sets ambitious targets for slowing pop-
ulation growth, calling for innovative measures in the
strengthening of reproductive health services and family
CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
planning and ensuring free access to information, educa-
tion, and contraceptive services.
REFORM STRATEGIES: COMMUNITY-BASED
HEALTH INSURANCE, PERFORMANCE-BASED
FINANCING, AND FISCAL DECENTRALIZATION
Rwanda has pioneered major programmatic, organiza-
tional, and health financing reforms, which are increasing
the accountability of major actors in the health sector.
Rwanda has a long history of centralized management
structures with a clear hierarchy and a relatively low level of
corruption. It has progressively moved toward a modern
health system design, including full autonomy of facilities,
decentralization, third-party payment, and strategic pur-
chasing through performance-based financing. It has tran-
sitioned from a faith-based service delivery model in the
colonial era to a model guided by the Bamako Initiative,
which sought to expand access to health services through
the development of local models of primary health care that
are managed and financed by communities. Both public and
private not-for-profit health facilities charge fees that are
locally retained and managed to cover the costs of health
services and improve the quality of care.
To improve financial access, the government pioneered a
microinsurance scheme and supported its expansion and sub-
sidization. It then introduced a mechanism of performance-
based financing to provide incentives to health facilities to
deliver high-impact interventions and ensure quality of ser-
vices. In 2006 it established a fiscal decentralization policy
and a legal framework that delineated a clear role for central
and local governments and service providers.
Together these three reforms constitute strategies to
strengthen accountability for services to citizens as part of
Rwanda’s 2006 national decentralized service delivery policy
(Government of Rwanda 2006). Fiscal decentralization has
been accompanied by measures to strengthen citizen partic-
ipation and accountability, including mechanisms for estab-
lishing accountability links between citizens and local
government officials, contractual performance between
health services providers and local governments or national
policy makers, and contractual approaches between com-
munities and health providers. This policy can be visualized
by using the accountability framework laid out in the World
Development Report 2004 (World Bank 2004). Accountabil-
ity of health providers to clients (“client’s power”) is strength-
ened through micro–health insurance funds that claim and
fund health services on behalf of households. Accounta-
bility of providers to the government (the “compact”) is
405
 strengthened thanks to performance-based financing mech-
anisms. Accountability of government to citizens (“voice”)
is strengthened through decentralization, citizen report
cards, and the possibility of recourse to the ombudsman
(figure 22.3).
Rwanda has finally settled into its current decentralized
model of care, in which health facilities are fully
autonomous entities responsible for the management of
financial resources, health service delivery, and human
resources for health. Community-based health insurance
schemes, which have been established and scaled up nation-
ally, have evolved in response to low levels of utilization of
health services. Partly to encourage community members to
buy in to these health mutuelle schemes, the government
developed performance-based financing as a complemen-
tary scheme to boost the performance and motivation of
staff to deliver higher-quality services as well as to increase
the delivery of preventive services.
Mutuelles and performance-based financing are two com-
plementary schemes. Both aim to shift health financing
mechanisms from inputs-based mechanisms toward output-
or results-based contractual mechanisms. The interactions
between the two strategies are therefore strong. Whereas
mutuelles emphasize personal curative care services, perfor-
mance-based financing emphasizes high-impact preventive
services and the quality of services.
The nationwide expansion of both strategies has
occurred rapidly since the pilot schemes were launched,
thanks largely to government ownership and commitment,
which have propelled these strategies forward. The nation-
wide implementation of mutuelles was closely followed by
the implementation of performance-based financing.
Mutuelles revealed the need for improvements in staff
motivation and incentives to deliver high-quality health
services, which in turn served as the driving force for per-
formance-based financing. As the implementation of
performance-based financing in the pilot districts proved
successful in improving staff motivation and health out-
comes, performance-based financing became a major pillar
within the Ministry of Health Strategic Plan (2005–09). In
2006 the government, with the support of external part-
ners, expanded performance-based financing to the entire
health sector (table 22.1).
Rwanda has continuously learned from and adapted
its health service delivery strategies. The ability of the
government to adapt strategies—as evident in the scaling
up of successful pilot schemes to the national level, in
light of the changing macroeconomic and health sector
environment—was also essential to strengthen health
services. Independent controls and quality checks under
contractual arrangements have been essential for the
monitoring and evaluation of health facility performance.
Performance-based financing is a prime example of a sys-
tem in which independent controls on performance are in
place to ensure proper monitoring and reporting of
health centers and district hospitals on the quantity and
quality of services delivered, which in turn drive facilities’
reimbursement.

Figure 22.3 Decentralization of and Accountability for Health Services in Rwanda

National government

Fiscal decentralization Performance-based

Citizen report Imihigo transfers
cards,
ombusdman
Local government
Compact
Voice

Client power

Clients/citizens Autonomous
facilities providers

Health insurance
(mutuelles)

Source: Government of Rwanda 2006.

Note: Imihigo are performance contracts in which the region and its districts promise the president of Rwanda that they will implement the measures
outlined in the annual plans.

406 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
 Table 22.1 Health Financing Reforms in Rwanda, 1999–2008
Reform 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
Enrollment
in mutuelles
(percent) 0 1.6 2.6 4.7 7 27 44 73 75 85
Number of
districts with
performance-
based financing 0 0 0 2 5 7 7 14 23 30
Assisted
deliveries
(percent) 39 52
Key policy Rwandaise Fiscal decentralization Health insurance
milestones d’Assurance law passed; law passed
Maladie (RAMA) decentralized
created; fiscal service delivery
decentralization policy includes
policy performance-based
financing and
mutuelles
Key Community-based RAMA Performance- Integration of Fiscal Full autonomy
implementation health insurance established based performance- decentralization given to public
milestones pilots launched financing based financing health facilities,
in three health pilots and subsidies to including over
districts since launched mutuelles in hiring and firing
July 1999 in four national budget National guarantee
districts fund established
Per capita health
expenditures
(National
Health
Account) $15 $37

Source: Authors.
Note: RAMA = Rwandaise Assurance Maladie.
407
 Community mobilization and intersectoral collabora-
tion have contributed significantly to the implementation of
health reforms. Cultural and social factors, particularly sol-
idarity within communities, have contributed to the success
of several health service delivery innovations. The rapid
proliferation of health mutuelles was made possible by the
strong solidarity within Rwandan society, in which commu-
nity members encouraged one another to join. This solidar-
ity has been long-standing in Rwanda, evident even before
the start of health mutuelles. In the case of health mutuelles,
cultural barriers initially thwarted implementation, particu-
larly because Rwandans were used to seeking care from tra-
ditional healers and because patients of traditional healers
were able to pay in kind rather than making cash payments.
Community-based health insurance
Health insurance has been scaled up at the national level on
the basis of community-based health insurance schemes,
with strong support from the government of Rwanda. Cov-
erage of health mutuelles increased dramatically between
2003 and 2008, rising from less than 7 percent to 85 percent
of the population in just five years.
Health mutuelles have had a significant positive impact
on health service utilization, income protection, and house-
hold health behaviors (Sekabaraga, Diop, and Soucat 2010).
Health insurance coverage increased the use of modern
health services by children under five between 2000 and
2007. In the general population, use of modern services by
the insured population is nearly twice as high as use by the
uninsured (figure 22.4). In addition, health mutuelles
seem to protect against expenditure caused by unexpected
illness, because uninsured households spend directly twice
as much as insured households for illness-related expendi-
ture. Health insurance coverage has thus significantly
reduced household out-of-pocket health expenditures
(Ministry of Health of Rwanda and World Bank 2011).
Among women who gave birth during the period 2000–05,
77 percent were affiliated with Rwandaise d’Assurance
Maladie (RAMA). About 42 percent of women affiliated
with a health mutuelle were assisted by a skilled health pro-
fessional during delivery compared with 35 percent of women
with no insurance (Sekabaraga, Diop, and Soucat 2010).
Before 1996 health services in Rwanda were free. Given
problems associated with the quality of services and the
financial sustainability of the health system, cost-recovery
mechanisms were reintroduced in 1996, while policy
debates continued over alternative financing mechanisms
for reconciling internal resource mobilization and access to
408 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
Figure 22.4 Use of Health Services in Rwanda in Event of
Illness, by Health Insurance Coverage, 2005
60
Use of health services (%)
50
17
40 9
30 20 9
13
11
20
11 35
25
10
10 16
0
None RAMA Health All
mutuelle
Public services Private services
Traditional practitioners
Source: Government of Rwanda 2005.
health services, including prepayment mechanisms through
health insurance schemes. In view of the significant scope of
poverty following the war of 1994, the reintroduction of
user fees in the health sector in 1996 was accompanied by an
exemption policy that allowed for free health care coverage
in health facilities for people identified by the local admin-
istrative authorities as indigent. Coverage of the poor and
vulnerable groups has been integrated in the development
of health mutuelles since they were first piloted in 1999.
In 2000 Rwandaise d’Assurance Maladie, the first health
insurance scheme for the formal sector, was established for
civil servants. Membership in RAMA soon became compul-
sory in order to increase coverage. To complement insur-
ance schemes such as RAMA and a few private insurance
schemes that target the formal sector, health mutuelles cov-
ering rural communities and the informal sector were
expanded to promote equitable access to quality health ser-
vices. Health mutuelles were designed to pool or spread the
financial risk of seeking care across their membership base.
The goal was to respond to the low use of health services
(caused in part by user fees) by improving financial access
to health services, particularly for underserved populations.
The package of services reimbursed by health mutuelles to
health facilities has expanded over time and currently cov-
ers all services delivered within a health center as well as
drugs from the national essential drug list. Health mutuelles
also cover most costs for health services and drugs delivered
at district and referral hospitals when mutuelle members
receive referrals for these higher levels of care.
These schemes have been extended to empower citizens
and communities in the health sector and to change their
interactions with health care providers. Through contrac-
tual relations between health mutuelles and health care
providers, communities and citizens have a tool with
which to hold health care providers accountable for ser-
vices provided. In essence, health mutuelles in Rwanda
reflected a bottom-up strategy, driven largely by local
communities. The success of the initial pilot schemes
motivated the central government to scale up this strategy.
Currently, the central government sets national guidelines
and policies, including benefit packages and contribution
policies. In 2008 it instituted a national guarantee fund for
providing general subsidies to support the extension of the
benefit packages of health mutuelles. Since 2009 the gov-
ernment has also built on the national network of health
mutuelles to elaborate and implement demand-based tar-
geted subsidies, through which the government, donors,
and nongovernmental organizations (NGOs) are providing
health insurance coverage to poor people, vulnerable
groups, and people living with HIV/AIDS. At the opera-
tional level, health mutuelles are run and organized by
community representatives and local health care providers.
They also serve as a forum for promoting dialogue between
the community and providers on the quality and range of
health services offered. In this way, community members
are better able to hold providers accountable for services
delivered.
For the majority of the population employed in the rural
and informal sectors, an incremental approach was followed
in developing mechanisms for pooling health risk. The
process of cumulative building of national capacities is a
hallmark of the incremental development of mutuelles in
Rwanda. Capacity building for local actors involved in set-
ting up, managing, and monitoring health mutuelles began
in 1999, with the establishment of mutuelles in three pilot
health districts.2 The geographic extension of the mutuelles
was accelerated in 2004, after the adoption of a national
strategic framework for their development. The number of
health mutuelles grew by a factor of 2.5 in a single year,
climbing to 226 nationwide. In 2007 each of Rwanda’s 403
health centers had a partner health mutuelle or “health
mutuelle section,” and all of the country’s 30 districts had a
district health mutuelle, which, on average, was linked to 13
health mutuelle sections.
The period of experimentation, which started in 2001,
was followed by attempts to adapt the institutional
arrangements for health mutuelles to the environment of
administrative and political decentralization and by early
efforts to expand the mutuelles to other districts of the
country. In the absence of an explicit policy framework for
CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
coordinating the initial efforts of adaptation and expan-
sion, a variety of local policies and incentives proliferated,
initiated by local authorities of all categories (political
groups, associations, and so forth) to motivate the popula-
tion to join health mutuelles (by, for example, linking
membership to civil status services, microcredit, and so
forth). These initiatives helped develop the health mutuelles
and contributed to the growth of social demand for their
expansion nationwide.
Current levels of contributions to health mutuelles are
affordable for all but the poorest 10 percent of Rwandans.
Affordability of enrollment in health mutuelles is assessed
based on the percentage of contributions in household total
expenditures and household nonfood expenditures. The
cost of membership rises with family size (for many reasons,
including high disease burden in large families and the
externalities of insurance benefits). Male-headed house-
holds have a higher proportion of enrollment than female-
headed households, partly because of the income difference
between the two groups. Family heads that completed pri-
mary school or received some vocational training tend
to have the highest rate of enrollment among the least-
educated and best-educated households. This trend is also
reflected in the enrollment rate by income (expenditure)
category, so that middle-income and middle-rich people
tend to have the highest participation rates in health
mutuelles. Out-of-pocket illness-related expenditures
among households enrolled in health mutuelles are twice as
high as those of members. Households that live very close to
health centers spend more than those living far away.
Households that are not covered by health mutuelles spent
nearly twice as much for illness-related services as people
who were insured.
Information has played an important role in the opera-
tional management and monitoring of the development of
health mutuelles at the local level and in their strategic man-
agement at the central level. Indeed, Rwanda is one of the
few African countries where an information system to sup-
port health mutuelle management has been developed to
permit monthly monitoring of performance. It is also one
of the few countries that has adapted training manuals on
health mutuelle development, management, and monitor-
ing in line with the local context, including availability in
the local language. Numerous other activities to promote
and raise awareness of health mutuelles are carried out
around the country and at the national level. The Ministry
of Health occasionally organizes an annual event on mutual
organizations at which prizes are awarded to the best per-
forming health mutuelles.
409
 Performance-based financing and reforms
of human resources management
Based on a positive evaluation of a three-year pilot phase in
two provinces, Rwanda has implemented a national pro-
gram since 2005; it has scaled up performance-based
financing since 2006 (Rusa et al. 2009).3 Performance-based
financing is currently implemented at three levels: health
centers, hospitals, and community levels.
The performance-based financing model is based on the
principle of separating purchaser and provider functions in
health service delivery (figure 22.5). By distinguishing
between and maintaining a split between bodies purchasing
services and bodies providing services, this model promotes
accountability and avoids conflicts of interest. The model
consists of a family of methods and approaches that aim,
through differing levels of intervention, to link incentives to
performance. In the national model for health centers,
payments for performance are based on the quantity of
outputs achieved conditional on the quality of services
delivered. Through performance-based financing, the central
government purchases 13 quantitative indicators and 13
qualitative measurements from health facilities (tables 22.2
and 22.3).4 At the hospital level, performance is assessed
through a peer-evaluation mechanism.
Low intake of preventive services and poor quality of care
in health centers served as the impetus for introducing per-
formance-based financing strategies. Quality of care became
a particularly salient issue as expansion of health mutuelles
increased utilization rates dramatically at all health centers,
adding to the workload of health personnel who, at the time,
had little or no incentive to take on the additional work. The
impetus for the performance-based financing strategy in
Rwanda first came from external actors; additional resources
and incentives were provided to health workers to improve
efficiency and outcomes, under pilot schemes implemented
in 2001–05. As a result of the pilots’ success, performance-
based financing became a major pillar within the Ministry of
Health strategic plan and was implemented nationally.
Rwanda’s institutional performance-based financing
model can be classified as “output-based financing,” because

Figure 22.5 Rwanda’s Performance-Based Health Care Financing Model

Purchaser-provider split in the national model

The Controllers
District Hospital (public, FBOs)
District Health Departments (USF)
Technical Assistance Agencies
Special Interest Groups
Steering Committee
Grass-root Organizations (piloted)

Rwanda
PBF

The Purchasers The Providers

GOR District Hospital (public, FBOs)
WB MAP Health Center (public, FBOs)
USG Second Tier:
(GF) by Health Centers
1. CHWs
2. TBAs
3. Private for-profit Providers
4. Private not-for-profit Providers

Source: Rusa et al. 2009.

Note: Purchasers are those who pay for services. They include NGOs, which purchase services with their own funds and act as fundholders or pass-through
mechanisms for other donors; the government of Rwanda (in the case of the national model); the U.S. government, through collaborative agencies such as
Management Science for Health and Family Health International; World Bank MAP funding, NGOs in the Cyangugu and Butare pilots and purchasing
performance through the new national model; and the Global Fund to Fight AIDS, Tuberculosis and Malaria, which is expected to start purchasing through
the new national model. Providers can include public and faith-based managed health facilities (health centers and hospitals) and private for-profit health
facilities. Controllers are those who control the level of performance, such as district health teams in the Ville de Kigali pilot, who certify a mix of quantity
and quality deliverables in health centers and hospitals. District health teams were used in Butare for random quantity control and in Cyangugu l for a qual-
ity measure in health centers. A peer-evaluation mechanism for district hospitals was piloted in Cyangugu and partially in Butare.

410 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
 Table 22.2 Output Indicators and Unit Payments under Performance-
Based Financing Formula
Amount paid
Output indicators per unit (US$)
Visit indicators (number of)
Curative care visits 0.18
First prenatal care visits 0.09
Women who completed 4 prenatal care visits 0.37
First time family planning visits (new contraceptive users) 1.83
Contraceptive resupply visits 0.18
Deliveries in the facility 4.59
Child (0–59 months) preventive care visits 0.18
Content of care (number of)
Women who received tetanus vaccine during prenatal care 0.46
Women who received malaria prophylaxis during prenatal care 0.46
At-risk pregnancies referred to hospital for delivery 1.83
Emergency transfers to hospital for obstetric care 4.59
Children who completed vaccinations (child preventive care) 0.92
Malnourished children referred for treatment 1.83
Other emergency referrals 1.83

Source: Basinga et al. 2009.

Table 22.3 Quality Indicators Services and Weights Used in the

Performance-Based Financing Formula
Share of weight Share of weight
allocated to allocated
structural to process Means of
Service Weight components components assessment
General administration 0.052 1.00 0.00 Direct observation
Cleanliness 0.028 1.00 0.00 Direct observation
Curative care 0.170 0.23 0.77 Medical record review
Delivery 0.130 0.40 0.60 Medical record review
Prenatal care 0.126 0.12 0.88 Direct observation
Family planning 0.114 0.22 0.78 Medical record review
Immunization 0.070 0.40 0.60 Direct observation
Growth monitoring 0.062 0.15 0.85 Direct observation
HIV services 0.090 1.00 0.00 Direct observation
Tuberculosis service 0.028 0.28 0.72 Direct observation
Laboratory 0.080 1.00 0.00 Direct observation
Pharmacy management 0.060 1.00 0.00 Direct observation
Financial management 0.050 1.00 0.00 Direct observation
Total 1.000

Source: Basinga et al. 2009.

it pays on a fee-for-service or case reimbursement basis to
improve outputs. Although performance-based financing
incentives are generally meant to induce providers (the sup-
ply side), supply-side incentives in Rwanda work through
supplier-induced demand, whereby suppliers actively seek
to convince people to use more of certain kinds of services.
Such incentives are necessary in Rwanda, where the goal is
not to limit excessive supply and unnecessary demand (as is
the case in richer health systems) but rather to induce
providers to provide more services while also increasing
financial revenues at the health facility level.
As of 2006 the government transferred about $1.80 per
capita from the Treasury directly to health facilities at the
basic health service level on the basis of a performance-
based formula. The program channels funds directly from
Treasury to the bank accounts of the more than 400 health
clinics in Rwanda (40 percent of them faith based, 60 percent
of them public) on the basis of performance agreements.

CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS 411
 These funds are flexible and may be used for facility expen-
ditures, including performance-linked salary bonuses,
partially substituting for revenues from user fees. Rwanda
implemented a three-tier performance-based financing
model—including hospitals, health centers, and ultimately
the community level—in order to make health services
more community oriented.
One of the key objectives of performance-based financing
was to introduce bonuses to health workers as incentives for
good performance, based on a range of agreed indicators.
This system was designed to allow for better monitoring of
health personnel activities and hence to enable district and
central levels to track staff performance over time. Although
the central government determines the overall performance-
based financing budget envelope that the health facility
receives, based on a formula involving the quantity and qual-
ity of services provided, it is the committee within the health
facility itself that determines how these funds should be used.
In 2008 Rwanda decentralized wages. As a result,
financing and payments for health personnel are increas-
ingly linked to performance in which block grants from
the government and donors can be used as salary. Direct
spending on wages and salaries by the central administra-
tion and transfers to public institutions for salaries of
health workers have declined. In contrast, funds chan-
neled to human resources for health through provinces
and districts that come from both the government
(including performance-based financing) and user fees
collected directly by facilities have risen dramatically in
recent years (figure 22.6).
Figure 22.6 Financing for Human Resources for Health
in Rwanda, 2005–08
25
22.4
20 18.5
US$ (millions)
15
11.2
10
8
6.7
5 3.0
0.5
0
05
06
07
08
20
20
20
20
Basic salaries Performance-based financing
Linear (basic salaries) Linear (performance-based financing)
Source: Authors.
412 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
It is the policy of the Ministry of Health, in collaboration
with development partners, to harmonize the framework
for compensation packages of health professionals. The
objective is to avoid introducing distortions in the distribu-
tion of health workers, which occurs when health workers
move from the public sector to donor projects where the pay
is higher. In light of this concern, donors, such as the Global
Fund, have begun to use national pay scales and to fully
integrate staff within the health system at large.
Results-based block grants in Rwanda have contributed
significantly to the increase in assisted birth deliveries as
well as the intake of child health services; the grants have
also increased the quality of services. Clinics that received
performance-based financing (of about $1.80 per capita
per year) performed more assisted deliveries and more
post-natal visits than clinics receiving the same funding
without a performance contract (figure 22.7). The quality
of care of antenatal services was 15 percent higher in per-
formance-based financing clinics than in control clinics
(figure 22.8).
The results achieved—in service supply and the enthusi-
astic participation of all stakeholders—after a few years of
experience point to a promising future. However, because it
is a dynamic strategy, performance-based financing adjusts
to innovative ideas that benefit the population and health
care providers. The remaining challenges are related to the
permanent oversight requirements, the accuracy of data,
and the delicate balance of the pricing of the various indi-
cators. The future of performance-based financing will
depend on finding appropriate solutions to these issues.
Figure 22.7 Number of Assisted Deliveries in
Rwanda under Performance-Based and
Nonperformance-Based Financing,
2006 and 2008
60
55.6
Number of assisted deliveries
7.3 % increase
49.7 due to PBF
50
8.2
40
36.3
34.9
30
Baseline (2006) Follow up (2008)
Control facilities Performance-based facilities
Source: Basinga et al. 2009.
Figure 22.8 Quality of Care in Rwanda under
Performance-Based and Nonperformance-
Based Financing, 2006 and 2008
0.20
0.15
0.15
0.10
Quality of care
15 % Standard deviation
0.05 increase due to PBF
0.00 0 autonomy to health facilities and transferred fiscal respon-
–0.05
–0.10 –0.10
–0.13
–0.15
Baseline (2006) Follow-up (2008)
Control facilities Performance-based facilities
Source: Basinga et al. 2009.
Fiscal decentralization
A strong commitment to bring services closer to the people
resulted in rapid fiscal decentralization, increased citizen
participation, and increased autonomy of health facilities.
Fiscal decentralization (adopted as policy in 2001 and
enacted into law in 2006) served as an essential component
of Rwanda’s decentralization agenda to devolve authority to
the district level. As was to be expected, given the presence
of such a strong state, a mindset change was needed to move
forward on many of these reforms. As in the case of decen-
tralization, some officials at the central level felt disempow-
ered and were initially unwilling to relinquish their control
at the outset. Officials at the local level had to adapt to
accept their new responsibilities, and donors had to adjust
to working with local governments.
The decentralization of authority across sectors was
planned through an incremental, three-phased approach.
The first phase (2001–05) focused on administrative and
political decentralization; it aimed to institutionalize
decentralized governance by establishing democratic and
community development structures, delineating policies,
establishing legal frameworks, and strengthening institu-
tional capacity at local levels. Phase two, which began in
2006 and ran through 2008, focused on making local gov-
ernments responsible for bringing health services closer to
beneficiaries. The devolution of responsibilities in health
services and the transfer of resources under fiscal decentral-
ization are the backbone of relationships between the
national government and districts in the health sector. It
aimed to reorganize roles and responsibilities within local
CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
government under the decentralization framework and
further strengthen district authority while allowing for
greater community participation and facilitating resource
allocation to local government. Central government respon-
sibilities in this phase remained regulation and development
of policy frameworks, capacity building of local govern-
ment, and monitoring and evaluation.
Finally, phase three, which began in 2008, granted
sibilities and financial resources from the central and local
government to facilities. This reform has resulted in relative
autonomy in budgeting and financial management within
facilities, because health care providers are now contracted
with and managed by health facilities.
Fiscal decentralization in Rwanda was government
owned and driven, with strong support and collaboration
of development partners. The objective was to bring ser-
vices closer to the people, and to improve the financial via-
bility of districts. The infrastructural changes needed may
not have been in place (until June 2007 districts lacked
accounting software to manage financial transactions, and
local capacity in managing financial and human resources
still remains limited), but overall success was striking. The
strategy was organized by the central government, which
from the outset determined the degree of authority delegated
to local levels and delineated relevant policies and standards.
In these efforts, the central government received significant
technical assistance and guidance from development part-
ners, which organized their support in the form of a sector-
wide policy to ensure government ownership over decision
making and policy setting; increase coherence between pol-
icy, spending, and actual results; reinforce the government’s
management systems; and harmonize donor support.
Decentralized units at local levels were given the author-
ity to manage the flow of funds (once received) as well as the
delivery of health services. Decentralization transformed
health facilities into autonomous entities, with the ability to
manage financial and human resources as they deem most
appropriate, according to local needs. The process gave
them complete control over the hiring and firing of health
personnel.
The accountability links between local governments
and national policymakers are strengthened through
inspections, audits, and Imihigo—performance contracts
in which the region and its districts promise the president
of Rwanda that they will implement the measures outlined
in the annual plans.5 Decentralization reforms have
resulted in increased responsibilities of local governments
in many areas and are increasing space for community
413
 participation and community-driven development initia-
tives. Satisfaction with service delivery is measured
through citizen report cards.
CONCLUSION
Fiscal decentralization, performance-based financing, and
the expansion of health insurance have led to a dramatic
increase in resources for frontline providers. Between 2002
and 2007, public resources flowing to health facilities more
than tripled (figure 22.9). The increase in resources took
place at all levels, showing no higher priority given to the
primary care level. An increasing number of donors are
channeling their assistance through on-budget support, and
major efforts are under way toward coordinating and har-
monizing aid. Most of the increase in publicly managed
resources flowed to human resources and to performance-
based financing; resources directly managed by donors
funded HIV/AIDS activities. On the domestic front, inter-
nally generated revenues of health facilities increased signif-
icantly as a result of increased utilization of health services
and health insurance coverage; an increasing share of inter-
nally generated revenues is captured by health centers,
strengthening frontline providers.
Rwanda chose to develop a mixed health care financing
model, combining decentralization and performance-based
financing with a strategy to pool private spending through
Figure 22.9 Financial Transfers to Districts for Priority Programs, 2006–10
9
8
6
Local currency
5 4.8
4 3.7
3 2.5
2
2 1.7
1
0.4 0.4
0
Community-based health
insurance
2006
Source: World Bank 2010.
414 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
the building of grassroots, community-based microinsur-
ance. Key reforms included provisions for financial protec-
tion and other support for poor people.
Lessons learned from the Rwandan experience provide a
strong base for future action. As financial barriers to access
to health services are being reduced significantly through
health care financing reforms, improving the quality and
sustainability of health services will remain among the main
challenges facing Rwanda in the coming years.
The success of Rwanda in improving health outcomes,
particularly for women, children, and the poor, can be
linked to both increased resources and implementation of
reforms. Rwanda used the inflow of resources to strengthen
its country system, including its public finance and health
systems. It designed its own brand of reforms, staying away
from donor fads and looking realistically and opportunisti-
cally at the balance between sustainability and equity.
Reforms focused on results, and results attracted more
funding, both domestic and external. Increased resources
and reforms mutually reinforced each other as part of a vir-
tuous cycle. Efficient and equitable use of resources required
reform, and the success of reforms needed resources.
The success of the health financing innovations is criti-
cally linked to the institutional context. Decentralization
reforms coupled with performance-based financing ensure
that health facility managers have not only the incentives
but also the power to ensure that these innovations translate
8.4
5.3
4
2.3 2.2 2.3
1.3
0.9
Performance-based Community health
financing
2007 2008 2009 2010
into changes in the delivery of services. The performance-
based financing system is now being extended to provide
incentives to community health workers providing outreach
services and demand-side incentives to women to continue
the increased utilization of key maternal health services.
Strong government leadership, vision, and the step-by-
step building of a policy and regulatory framework at all
levels have fostered the short- and long-term sustainability
of health sector reforms. The Rwandan government
showed flexibility and was able to adapt strategies in light
of the changing macroeconomic and international health
environment. Government coordination of donor funding
was critical to ensure that aid was used effectively and
aligned with national priorities. Systems for improved
accountability, including contractual arrangements, inde-
pendent controls, and quality checks, were essential for
monitoring and evaluating health facility performance.
Cultural and social factors, particularly solidarity within
communities, also contributed to the success of several
health service delivery innovations.
A particularly important feature of the Rwanda experi-
ence is the integration of strong mechanisms to evaluate the
impact of its policies. Rwanda’s experience, as well as that of
other low-income countries, in introducing pro-poor
financing policies needs to be systematically evaluated using
rigorous metrics and standardized benchmarks. It is possi-
ble for countries to nest impact evaluation designs when
introducing new policies at scale. The donor community
needs to support these evaluations if it wants to improve aid
effectiveness.
A key issue for the future is sustainability and the neces-
sary evolution of the institutional support for the health
financing approach of Rwanda. Like many Sub-Saharan
African countries, Rwanda is highly dependent on aid and
will remain so for the next decade. Sustained support from
the donor community is therefore needed to support the
health system strengthening agenda. Mutuelles provide one
way to ensure more sustained domestic funding. They rep-
resent an efficient way to pool private out-of-pocket spend-
ing, but there is a need for the government—with the help
of donors—to subsidize the enrollment of the poorest
Rwandans and to regulate the package of benefits as well as
the provider payments mechanisms to ensure equitable
access to quality services.
NOTES
1. Comprehensive knowledge of HIV/AIDS, which can be
used as a tracer indicator of general health knowledge in the
CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
country, means knowing that use of condoms and having a
single, uninfected, faithful partner can reduce the chances
of contracting HIV, knowing that a healthy-looking person
can have HIV/AIDS, and rejecting the two most common
local misconceptions about HIV/AIDS transmission and
prevention.
2. At the time, the districts were called Byumba, Kabgayi,
and Kabutare. They are now called Gicumbi, Muhanga, and
Save.
3. This program has been supported by a broad consor-
tium of donors, including the World Bank, the U.K.
Department for International Development, the European
Union, Sweden, the African Development Bank, the
Netherlands, and Germany.
4. A separate contract channels earmarked funds of global
HIV/AID programs for another 16 indicators.
5. Imihigo are also monitoring instruments designed to
help local authorities plan and act realistically.
REFERENCES
Basinga, P., P. Gertler, A. Binagwaho, A. Soucat, J. Sturdy, and
C. Vermeersch. 2010. Paying Primary Health Care Centers
for Performance in Rwanda. World Bank, Washington, DC.
Government of Rwanda. 2005. “Enquete Integrale sur les
Conditions des Menages.” Kigali.
———. 2006. Decentralized Service Delivery. Kigali.
May, J. F., and A. Kamurase. 2009. Demographic Growth and
Development Prospects in Rwanda: Implications for the
World Bank. World Bank, Washington, DC.
Measure DHS and ICF Macro. 1995. Demographic and
Health Survey, Rwanda. Calverton, MD. http://www
.measuredhs.com/.
———. 2000. Demographic and Health Survey, Rwanda.
Calverton, MD. http://www.measuredhs.com/.
———. 2005. Demographic and Health Survey, Rwanda.
Calverton, MD. http://www.measuredhs.com/.
———. 2007. Demographic and Health Survey, Rwanda.
Calverton, MD. http://www.measuredhs.com/.
———. 2008. Demographic and Health Survey, Rwanda.
Calverton, MD. http://www.measuredhs.com/.
Ministry of Health of Rwanda, and World Bank, 2011.
Country Status Report on Health, Health Systems and
Poverty. Human Development Series. World Bank,
Washington, DC.
Rusa, Louis, Miriam Schneidman, Gyuri Fritsche, and
Laurent Musango. 2009. “Rwanda: Performance-Based
Financing in the Public Sector” In Performance Incen-
tives for Global Health: Potential and Pitfalls, eds. Rena
Eichler, Ruth Levine, and the Performance-Based
415
 Incentive Working Group. Washington, DC: Center for
Global Development.
Sayinzoga, K., A. Soucat, A. Kjellgren, and P. Musafiri. 2010.
From Reconstruction to Development: Budgeting for Per-
formance in Rwanda. Human Development Series, World
Bank, Africa Human Development, Washington, DC.
Sekabaraga, C., F. Diop, and A. Soucat. 2010 “Can Innova-
tive Health Financing Policies Turn the Tide on the
Health MDGs? Evidence from Rwanda.” Communica-
tion to the International Symposium on Health Systems
Research, Montreux, Switzerland.
World Bank, 2004. World Development Report 2004: Making
Services Work for Poor People. Washington, DC: World
Bank.
———. 2010. Country Status Report on Health, Health
Systems and Poverty. Washington, DC: World Bank.

416 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS
 CHAPTER 23
The Malaria Control Success Story
Anne-Maryse Pierre-Louis, Jumana Qamruddin,
Isabel Espinosa, and Shilpa Challa
alaria is both preventable and treatable. Yet
M
more than 220 million cases of malaria are esti-
mated to occur each year, and approximately
785,000 people die from the disease annually. Half of the
world’s population—some 3.3 billion people living in 109
countries—are at risk of malaria (Roll Back Malaria 2008).
Worldwide, malaria is the fifth-leading cause of death
from infectious diseases (after respiratory infections,
HIV/AIDS, diarrheal diseases, and tuberculosis). The dis-
ease is life threatening and needs early, accurate diagnosis
and treatment, which can be difficult in remote areas that
lack clinics, trained health care providers, technical assis-
tance, or medicine.
Malaria is caused by Plasmodium parasites, which are
spread to humans through infected Anopheles mosquitoes,
called malaria vectors, which bite mainly between dusk and
dawn.1 Ninety percent of the world’s malaria deaths occur
in Sub-Saharan Africa, where the most severe form of the
disease prevails, making malaria the second-leading cause of
death in the region after HIV/AIDS (Roll Back Malaria
2008). The disease is one of the leading causes of death of
children under the age of five in Sub-Saharan Africa and a
major cause of complications, including maternal death and
low birth weight, in pregnancy.
Malaria is not only a health problem in Africa; it is also a
development problem. Death and disability (both short-
and long-term) from malaria have enormous social and
economic costs, costing African countries an estimated $12
billion a year in lost productivity (Gallup and Sachs 2001).
Treatment of severe episodes, which can cost up to one-
quarter of a household’s monthly income, accounts for up
to 40 percent of public sector health expenditures in the
worst-affected countries (WHO 2007). Other direct costs
include the costs of insecticide-treated nets and indoor
residual house spraying to prevent malaria, antimalarial
drugs, transportation to clinics, and hospital fees. Operating
in a vicious cycle, malaria is both a cause and a consequence
of poverty.
Malaria keeps countries as well as households in poverty:
annual economic growth in countries with high malaria
transmission has historically been lower than in countries
without malaria. Economists have estimated that malaria is
responsible for an “economic growth penalty” of up to 1.3
percent a year in malaria-endemic African countries (Gallup
and Sachs 2001). The disease discourages local and for-
eign investment and tourism, affects land use patterns and
crop selection (resulting in suboptimal agricultural produc-
tion), and reduces labor productivity through lost work days
and diminished on-the-job performance. It affects learning
and academic achievement through frequent absenteeism of
teachers and students. In children who suffer severe or fre-
quent infections, it can cause cognitive impairment and in
some cases permanent neurological damage. Other indirect
costs include loss of income and work, including the unpaid
work carried out largely by women who take care of the sick
and support them at home or in the hospital.
Expenditures on malaria control drain already fragile
economies through the deterrent effect on investment
417
 (private businesses are reluctant to expand in areas where
the disease is affecting the workforce and creating gaps in
the production line). Economic and social decision mak-
ing at all levels is affected. Travelers are hesitant to visit
countries with a high incidence of malaria, reducing
tourism revenues and leading to losses of job opportuni-
ties and income.
A global effort to help countries control the disease
that was revitalized in 2005 is showing signs of success. In
countries in which control measures have been intensi-
fied, there have been clear and positive results. Eleven
countries and one territory in Africa show reductions of
more than 50 percent in either confirmed malaria cases or
malaria admissions and deaths in recent years (WHO
2010b). This progress is a direct result of the scaling up
and acceleration of measures against the disease. Other
positive outcomes include parallel declines in child mor-
tality in some countries.
This chapter examines progress in Sub-Saharan African
countries on key malaria indicators as well as in funding for
the fight against malaria between 2005 and 2010. The first
section examines the four main interventions for preventing
and treating malaria (insecticide-treated nets, antimalarials,
artemisinin-based combination therapy, indoor residual
spraying). The second section looks at prevention among
and treatment of pregnant women and children under five.
The third section discusses partnership and coordination in
the fight against malaria. The fourth section describes suc-
cess stories in four African countries (Eritrea, Ethiopia,
Rwanda, and Zambia). The last section identifies lessons
learned and discusses the way forward.
INTERVENTIONS FOR PREVENTING AND
TREATING MALARIA
Between 2000 and 2008 malaria control efforts in Africa
were measured against targets set in the 2000 Abuja Decla-
ration.2 Since 2008 progress has been measured against the
universal coverage targets set forth in the Roll Back Malaria
Global Malaria Action Plan 2010. These targets include the
following:
■ Eighty percent of people at risk from malaria are using
locally appropriate vector control methods, such as
long-lasting insecticidal nets, indoor residual spraying,
and, in some settings, other environmental and biologi-
cal measures.
■ Eighty percent of malaria patients are diagnosed and
treated with effective antimalarial treatments; in areas of
418 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY
high transmission, 100 percent of pregnant women
receive intermittent preventive treatment.
■ The global malaria burden is reduced by 50 percent from
2000 levels, to less than 175 million–250 million cases
and 500,000 deaths annually.
This section looks at progress on key malaria indictors in
Africa based on an analysis of nationally representative sur-
veys in countries in which the data were available for a par-
ticular indicator. A subset of countries for which data were
available is used to illustrate progress on key indicators.
Insecticide-treated nets
The promotion and distribution of insecticide-treated nets
has long been recognized internationally as a key cost-effec-
tive intervention to control malaria (Statesmen’s Forum
2010 (box 23.1). This intervention has been the primary
focus for scale-up over the past few years. Beginning in
2008, some countries in Africa made a policy shift from pro-
viding insecticide-treated-net coverage only for populations
at greatest risk (including children under five and pregnant
women) to seeking coverage for the entire population at
risk. This shift has largely been the result of a global effort to
increase progress toward the Millennium Development
Goals (MDGs) and a call for action from the UN Secretary
General in 2008. Given that the policy changed fairly
recently, the shift to implementation has not yet been seen
on a large scale across the continent.
Figure 23.1 illustrates progress toward coverage with
insecticide-treated nets across a number of countries in
Africa. At first glance, the story appears discouraging, as
many countries are well below the 80 percent coverage
target for this key intervention. Many of the data used,
however, come from representative surveys in countries
conducted in 2006 and 2007—around the same time that
malaria control efforts on the continent started ramping up.
Among countries in which coverage is 50 percent or greater
(Ethiopia, Madagascar, Rwanda, Senegal, Zambia, and
Zanzibar), most surveys were conducted between 2007 and
2009. The most recent 2010 Malaria Indicator Survey in
Zambia (not illustrated in the figure) shows an increase in
the share of households with at least one insecticide-treated
net rising to 64 percent.
Since 2006 coordinated efforts have supported the
“catch-up, keep up” approach to bed net distribution in
countries with low coverage. During the “catch-up” phase,
bed nets are distributed through either stand-alone or
integrated bed net campaigns to increase coverage
Box 23.1 Cost-Effective Tools in the Fight against Malaria
A variety of antimalarial tools are starting to pro-
duce results in Africa. They include artemisinin-
based combination therapies (ACTs), long-lasting
insecticidal-treated nets, indoor residual spraying,
intermittent preventive treatment, and rapid diag-
nostic tests.
Artemisinin-based combination therapies
Artemisinin-based combination therapies are pro-
duced by combining compounds from the Artemisia
annua plant with various antimalarial partner drugs.
They are recognized by the World Health Organiza-
tion (WHO) and the scientific community as the
most effective therapy for treating Plasmodium falci-
parum malaria, the predominant form in Africa and
the most deadly. ACTs are thought to delay the devel-
opment of drug-resistant strains of the disease. They
are also active against gametocytes, the sexual stage of
the parasite cycle, effectively reducing disease trans-
mission. For these reasons ACTs have become the first
line of therapy for many countries struggling to
control malaria infection and transmission. The
treatment is fast acting and produces minimal side
effects, making it possible for patients to return
quickly to their daily routines. The challenge is to
ensure that prompt and effective treatment with ACT
is available as a critical intervention to complement
prevention activities.
Long-lasting insecticidal nets
WHO policy recommends long-lasting insecticidal
mosquito nets (a type of insecticide-treated net),
which offer protection against malaria. If used by at
least 80 percent, in an affected area, these nets can
help break the malaria transmission cycle, thus
reducing the risk for all who live nearby. Woven from
insecticide-bound fibers, the nets kill the potential
disease-carrying mosquitoes on contact and offer
protection from bites. Unlike insecticide-treated
nets that require redipping every few months, long-
lasting insecticidal nets remain effective for at least
three years. The nets provide protection at night,
when individuals are most vulnerable. They are safe
for children and simple to hang up.
Indoor residual spraying
Indoor residual spraying is the application of long-
acting chemical insecticides on the walls and roofs of
all houses and domestic animal shelters in a given area
to kill the adult vector mosquitoes that land and rest on
these surfaces. Indoor residual spraying shortens the life
span of vector mosquitoes so that they can no longer
transmit malaria parasites from one person to another,
and reduces the density of the vector mosquitoes.
Intermittent preventive treatment
Intermittent preventive treatment given to pregnant
women at routine antenatal care visits has been shown
to promote healthier pregnancies and yield benefits for
both mothers and developing fetuses. It can signifi-
cantly reduce the proportion of low birthweight infants
and maternal anemia. Intermittent preventive treat-
ment in infants may also play a major role in malaria
prevention at the public health level; however, addi-
tional research is needed to confirm this hypothesis.
Rapid diagnostic tests
Rapid diagnostic tests for malaria, also known as dip-
sticks or malaria rapid diagnostic devices, have the
potential to greatly improve the quality of management
of malaria infections when high-quality microscopy is
not readily available. The tests detect specific antigens
(proteins) produced by malaria parasites that are pres-
ent in the blood of infected individuals. Some rapid
diagnostic tests detect only one species (P. falciparum),
others also detect other species of the parasite. Blood for
the test is commonly obtained from a finger prick.
WHO (2010) now recommends that all cases of sus-
pected malaria be confirmed with a diagnostic test
before treatment. As the incidence of malaria continues
to decrease in Africa, the need to differentiate between
malaria from nonmalarial fevers becomes more pressing,
especially given the need to avoid drug resistance and
increase efficiency of resources. Implementation of the
WHO recommendations will require a more integrated
approach to childhood illnesses (to address the question
of what to do if the fever is not caused by malaria).
CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 419
 Figure 23.1 Coverage of Insecticide-Treated Nets in Africa, by Country

100
90
Percent of households with at least

80
one insecticide-treated net

70
60
50
40
30
20
10
0

IS 7 a
Bu N on M ICS 06

o HS 06
a F ia S 2 06

6
go l G 06

ag da IS 6
6
f C ria S 7

ba o D S 2 6

er a M S 2 6

Sie rin wi M IS 2 a

ui N o M S 2 6
Re nd IS 06

au S 06

e IS 6

AI IS 08
nz iop IS 08
Za gal S 20 8
ni CS 07

m MIS 08

08
a- ger ICS 08
ric U tan HS 07
M bw H 008

S 06

bi eny CS 06
aL e S 7

Se ar D S 2 7
Et li D S 2 9
bl H 07

d ala M uine
ua H 00

0
as DH 00
. o ige IC 00
Zi on DH 200

am an IC 0

Th M 200

ne H 00
rr cip IC 00

c H 0
M IC 200
0

M 0
0
20

An oria S 20

20
C Gh c M S 20

To e D 20
an ga ia M 20

ss H 0

ar M 20
Za Eth ia M 20
Be MI S 20

20
0
ri D 0

eo MIC 20

am K MI 20

as D 20
pu a D 20

S/ 20
Eq n D 2
ep N re M S 2

Bi D 2
au e S 2

C
H

ad n M
I
Ivo D

M
an M la

M Rwa pia
a, a

ib ia
d’ d

b
a
t
e lan

n
g

in ib

g
i

o
rk am

ne i

hi
ôt zi
C Swa

G
G
.R

é
Af

m
em

To
al
tr
D

o
en

Sã
C

Source: WHO 2010b.

Note: DHS is Demographic Health Survey. MICS is Multiple Indicator Cluster Survey. MIS is Malaria Indicator Survey. AIS is AIDS Indicator Survey. Data
for Equatorial Guinea are from a national survey.
a. Updated since original DHS.

quickly. During the “keep-up” phase, bed nets are distrib- USE OF INSECTICIDE-TREATED NETS AMONG
uted through mechanisms such as routine antenatal care VULNERABLE GROUPS
visits, child wellness weeks, routine immunization com-
Two of the many vulnerable population groups in the malaria
pletion, and school health programs. This approach has
epidemic are pregnant women and children under age five. To
been proven to work in countries such as Ethiopia, Zam-
help safeguard these populations, a number of countries in
bia, and Zanzibar, where high coverage of bed nets is cor-
Africa made the strategic choice to focus key prevention and
related with sharp declines in malaria prevalence. Other
promotion activities on these two groups. Despite the recent
countries that are lagging behind, such as the Democratic
policy shift to universal coverage, many countries are still tar-
Republic of Congo (where just 4 percent of the popula-
geting these groups, given constrained budget envelopes for
tion has bed nets) and Nigeria (3 percent), have been
health in general and for malaria in particular. Examining use
focusing efforts on increasing coverage of bed nets. The
of LLINs among pregnant women and children under five
Democratic Republic of Congo plans to launching a
thus still provides a good indicator of progress on use of bed
large-scale bed net campaign in 2011 to ramp up coverage
nets (figure 23.2). Among these groups, the rates of usage are
rates. In Nigeria the government has been working with
more encouraging, with Equatorial Guinea, Ethiopia, The
partners to implement a “catch-up, keep-up” approach to
Gambia, São Tomé and Principe, Zambia, and Zanzibar
increasing bed net coverage. The catch-up phase has been
reaching at least 40 percent usage.
under way since 2008, through a phased-in nationwide
Analysis of usage rates among pregnant women indicates
mass campaign whose ultimate goal is to distribute
that, despite some progress, there is much room for
70 million nets (providing every household with at least
improvement. Figure 23.2 highlights two key issues in this
two long-lasting insecticidal nets) by the end of 2011. In
context: the lack of data for this indicator and the relatively
Nigerian states in which the bed net campaigns have been
low level of use among pregnant women compared with
completed, data from a Lot Quality Assurance Sampling
children under five. Ethiopia and Zambia4 are the highest
(LQAS) survey conducted in 2010 show coverage of bed
performers with 43 percent of pregnant women sleeping
nets of more than 70 percent.3
under an insecticide-treated net.

420 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY

Figure 23.2 Use of Insecticide-Treated Nets in Africa by Children under Five and Pregnant Women, by Country

100

90
Percent of households with at least

80
one insecticide-treated net

70

60

50

40

30

20

10

0
Zi ong HS 06

M bwe HS 8

ric U tani S 2 7
Re nd IS 6
bl H 07

am an IC 06
rk oo IC 06
as ICS 06

to S 6

o G 6
aw M ea

rr cip CS 07
eo MIC 006

go HS 06
8
bi au HS 06
e S2 6
M ICS 06

Rw iopi HS 6
d IS 6
ne HS 07

nz pia IS 8
AI S 2 8
M 7a
08
0

H 00
an ga a M 00

Be MIC 200

ua H 00
An rial 200

ne Nig MIC 200

Th IC 00

0
an M 00

Za al M 200

Za thio ia M 00

0
an Mal la uin
D 0
ba D 20

pu a D 20

0
Bu er a M S 20
a F M 20

Sie rin i MI 20

To D 20

0
Et li D 20

Se a D 20

ib M 20

20
S/ 00
f C ia S 2

G ic M S 2

Eq n D S 2

G a-Bi er D S 2

M 2

b IS 2
in n S

IS

ne S

IS
. o iger IC

ar I
M

a
m o,

g
ire

aL e

a
ni

m
g

am ss
Ivo

ri

h
h
au

a,
N
d’

E
e

d
C
ôt

ui
ep
C

G
.R

é
Af

m
em

To
al
tr
D

o
en

Sã
C

% of HH with >1 ITN % of children <5 years who slept under an ITN
% of pregnant women who slept under an ITN

Source: WHO 2010b.

Note: DHS is Demographic Health Survey. MICS is Multiple Indicator Cluster Survey. MIS is Malaria Indicator Survey. AIS is AIDS Indicator Survey. Data
for Equatorial Guinea are from a national survey.
a. Updated since original DHS.

Antimalarials use combination therapies, preferably those containing

artemisinin derivatives for p. falciparum malaria. In 2006 it
Fever is internationally recognized as one of the main indi-
developed standard guidance on artemisinin-based combi-
cators of malaria. Prompt access to treatment of malaria,
nation therapy for countries to adopt into policy. As of
especially among children under five, is one of the key indi-
2010, 43 countries in Africa had adopted ACTs, several as
cators of effective malaria treatment.5 The percentage of
first-line treatment and a few as second-line treatment. The
children under five receiving any antimalarial treatment
rollout of this policy is being implemented in stages across
(including artemisinin-based combination therapy, or
Africa (table 23.1).
ACT) through public and private channels varies by coun-
ACTs retail for $6–$10 a course in the private sector,
try (figure 23.3), but across the board there is vast room for
making the drug prohibitively expensive; even the $1 paid
improvement. In several countries, access to antimalarial
by the public and nonprofit is 10 times the price of
treatment is much higher in the public sector than in the
cholorquine. In addition, a number of factors, including
private sector, although many seek treatment first in the pri-
procurement and supply chain issues, theft, and counter-
vate and informal sectors.
feiting have made the availability of quality-assured
ACTs through public and private sectors challenging
Artemisinin-based combination therapy
(figure 23.4). But innovative solutions to access to ACTs
In response to increasing levels of antimalarial resistance, in are being implemented at the global and country level.
2001 WHO began recommending that countries experienc- The Affordable Medicines Facility for Malaria mechanism
ing resistance to conventional monotherapies, such as aims to reduce the unit cost of ACTs (box 23.2). In Zambia,
chloroquine, amodiaquine, or sulfadoxine-pyrimethamine, a supply chain intervention has dramatically increased the

CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 421

 Figure 23.3 Percentage of Children under Five with Fever Receiving Any Antimalarial Medication in the Public and
Private Sector, by Country

50

45

40

35

30

25

20

15

10

0
ila H 06

S 8
oz er H 07

rr que S 5
Zi Leo H 008
bw H 07

S 8
05

N ad IS 2 7
ib H 07

w S 4
. o eg HS 7
on H 04

ny HS 5

7
Be DH 008

rk han H 06

U so D S 2 6
nd HS 08

ne HS 03

S 6
oo H 05

Ta bia S 2 7
an HS 4

S 7
04
hi D 00

Sie mbi DH 200

on H 0

C a M 200

ala H 0
ep Sen i D 200

Ke o D 200

N DH 200

a H 0

Li a D 200

H 0
nz D 00
D 200
az r D 20

M Nig a D 20

ba D 0
C e D S 20
Rw o D 20

am D 0
M ia D S 20

f C al D 20

Bu G in D 20
a F D 20

0
G a D 20

er a D 20
Za n D S 20

20
Et nd S 2

D 2
m ne S 2

r S2
Sw sca HS

ia S

d S

S
in a S
op H

C be H

H
ag i D

ia
ad Mal

ige
i

am ri
an
g

n
h

m
ui
ga
a

a
a
M

.R
em
D

Public sector Private sector

Source: WHO 2010b.

Note: DHS is Demographic Health Survey. MIS is Malaria Indicator Survey. AIS is AIDS Indicator Survey.

availability of ACTs and other essential drugs in health

Table 23.1 Adoption of Policies for Malaria
Treatment in Africa facilities (box 23.3).

Policy
Number of endemic countries and territories
Number of P. falciparum endemic countries
and territories
Artemisinin-based combination therapy
(ACT) used for treatment of P. falciparum
ACT provided free of charge in public
sector for all age groups
ACT provided free of charge in public
sector only for children under five
ACT delivered at community level
Pre-referral treatment with
quinine/artemether IM
Artesunate suppositories
Source: WHO 2010.
Number
Indoor residual spraying
43
Indoor residual spraying is recommended for control of
42 malaria in 71 countries, 32 of them in Africa. It is the pri-
42 mary vector control intervention in Botswana, Mozam-
bique, Namibia, South Africa, Swaziland, and Zimbabwe.
24 The number of people benefiting from indoor residual
5
spraying rose from 10 million in 2005 to 73 million in 2009,
25 or about 10 percent of the population at risk for malaria
(WHO 2010b). In countries that have seen a rapid decline
32
in malaria within the past five years, bed net distribution has
25
been complemented with indoor residual spraying in the
areas of the countries where it is epidemiologically justified.

422 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY

Figure 23.4 Percentage of Children under Five with The data also point to deficiencies in the use of intermit-
Fever Receiving Artemisinin-Based tent preventive treatment of pregnant women. In three
Combination Therapy in the Public and countries (Malawi, Senegal, and Zambia), at least 70 percent
Private Sector, in Selected Countries of pregnant women reported having received SP/Fansidar
during an antenatal care visit (figure 23.5).6 In contrast, less
14
than 10 percent of pregnant women in Benin, Niger, and
12 Nigeria and less than 20 percent of women in the Demo-
10 cratic Republic of Congo reported receiving intermittent
preventive treatment during an antenatal care visit.
8

6
PARTNERSHIP AND COORDINATION
4
FOR RESULTS
2
During the past decade governments, communities, donors,
0 and individuals have been working under growing consen-
06

20 o

08

08

07

06

07
S ng

sus and coordinated action. Since 2005 partners and

20

20

20

20

20

20
07
H o
D fC
S

S
H

H
malaria-endemic countries have placed greater emphasis on
.o
D

aD

aD

aD
ep
n

na

ria
ny

nd

bi
ni

strengthening coordination of technical and human

.R

ha

be

m
Be

Ke

ga
G
em

Za
Li

resources at global, regional, and country levels. This

D

Public sector Private sector
Source: WHO 2010b.
Note: DHS is Demographic Health Survey.
Box 23.2 The Affordable Medicines Facility
for Malaria
Most malaria cases are treated with an over-the-
counter drug such as chloroquine. Resistance to this
drug has increased, however, making it increasingly
ineffective. As a result, in 2004 WHO recommended
that all countries use ACTs as first-line treatment for
malaria. ACTs are not widely available at private phar-
macies and other medical distribution centers
because of their high cost.
The Affordable Medicines Facility for Malaria
(AMFM) is a financing mechanism, managed by the
Global Fund to Fight AIDS, Tuberculosis and Malaria,
designed to subsidize ACTs. The initiative negotiates a
lower price with pharmaceutical producers. It aims to
make ACTs even more affordable through a co-
payment and subsidy system that will allow first-line
buyers, including in-country private sector whole-
salers, hospitals, and nongovernmental organizations,
to pass on the benefit to patients, who will pay
$0.20–$0.50 for ACTs—comparable to what they
currently pay for less effective alternatives.
Source: Authors.
intense collaboration, under the umbrella of the Roll Back
Malaria Partnership (box 23.4), has been a key component
of the success on the ground in Africa. These partnerships
have helped reduce duplication in efforts, increase resource
mobilization, raise awareness of the problem at the global
and regional level, and create a network of technical and
implementation experts for countries to draw on as they
implement strategies to control malaria.
Funding for malaria control has increased dramatically
since 2005 (figure 23.6). Evidence-based advocacy efforts
have led to a number of notable achievements, including
increased funding from partners such as the President’s
Malaria Initiative, the World Bank, and the Global Fund to
Fight AIDS, Tuberculosis and Malaria. As a direct result of
these efforts, in 2008 the Global Fund increased its funding
to malaria to the historic level of $1.62 billion worldwide. In
addition, the Roll Back Malaria Partnership formed the Har-
monization Working Group, which seeks to harmonize
donor resources to support countries with costed national
malaria control plans. Partners also joined forces to develop
the Global Malaria Action Plan, which provides a framework
for current and future malaria control effort worldwide.
Aid harmonization for malaria has improved, and
malaria disbursements rose to $1.5 billion in 2009. But new
commitments for malaria control plateaued in 2010 at $1.8
billion. Government spending for health in high-burden
countries remains low, with far less than 15 percent of budg-
ets spent on the health sector in more than 90 percent of
countries. The level of spending for malaria falls far short of
the resources needed to control the disease, which the RBM

CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 423

 Box 23.3 Improving Access to ACTs in Zambia
Recognizing that access to proper treatment needed to
be improved, the World Bank, the United Kingdom’s
Department for International Development, and the
U.S. Agency for International Development (USAID)
joined forces to try out new drug distribution methods
in 16 districts in Zambia. The results from this pilot,
evaluated through a rigorous impact evaluation, have
been exceptionally encouraging. Simple but smart
steps to grease the supply chain for lifesaving drugs—
including hiring district-level planners to help manage
orders and deliver them more efficiently—have proved
very effective.
Two models were tested. Under Model A drug stock
continues to be held at the district level. The new posi-
tion of commodity planner (CP) at the district level,
designed to enhance planning capacity, is responsible
for coordinating orders from the health facilities and
stock management at the district. The CP also ensures
that requisitions requests are sent every month by each
health facility to the district store and performs picking
and packing operations at the district level to fulfill the
order requisitions of health facilities under that dis-
trict. The CP also estimates the overall requirements
and places orders to Medical Stores Limited (MSL)1 for
the stock needed to maintain the desired inventory
level at the district store. CPs were recruited to hold the
logistics responsibilities in 12 of the intervention
Figure B23.3 Baseline and Endline Stock-Out Rates following Implementation of Pilot Program
a. Districts implementing Model A
SP (malaria prevention)
46
DepoProveraa (contraceptive) 38
17
CTXa (antibiotics) 42
37
Amoxicillin (antibiotics)
31
ACT adulta (malaria treatment) 43
22
ACT pediatric (malaria treatment) 34
30
0 10 20 30 40 50 60 70 80
Percent
Source: Authors.
Note: Figures show the percentage of health facilities that were out of stock of the drug indicated. The intervention had similar
impacts on almost all essential drugs tracked.
a. Reduction in stock-out rate is statistically insignificant compared with observed change in control districts.
424 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY
districts, while the remaining 4 districts relied on phar-
macy technologists.
Model B eliminates the intermediate storage of drugs
at the district level. The district store is converted into a
“cross-dock,” a point of transit that receives shipments
from MSL that are pre-packed for individual health facil-
ities. Under this option, the district does not carry any
stock or perform any secondary picking and packing.
This system has the potential to reduce the scope for pil-
ferage and leakages because it enables better shipment
tracking.2 However, it hinges on the health facilities
transmitting their order requisitions to MSL every
month to allow for the assembling of packages to indi-
vidual health facilities. As in Model A, a CP was added to
the district store under this option to ensure the smooth
flow of order information from the health facilities to
MSL.
Remarkable improvement in access to essential drugs
has occurred at the health facility level, particularly in dis-
tricts in which Model B was implemented. In districts
implementing Model A (panel a of figure B23.3), 38 per-
cent of health facilities were out of stock of DepoProvera
(a contraceptive) at baseline. At the end of the pilot pro-
gram, the stock-out rate had fallen to 17 percent. Reduc-
tions in stock-out rates of the same magnitude are
observed for amoxicillin and ACT for adults. Reductions
in stock-out rates were even more dramatic in districts
b. Districts implementing Model B
58 SPa (malaria prevention) 52
16
DepoProveraa (contraceptive) 1 45
CTXa (antibiotics) 40
33
74 Amoxicillina (antibiotics) 72
16
ACT adulta (malaria treatment) 48
6
ACT pediatrica (malaria treatment) 43
12
0 10 20 30 40 50 60 70 80
Percent
Baseline Endline
 Box 23.3 (continued)

implementing Model B (panel b of figure B23.3), where and over-five mortality by 25 percent. These reductions
decreases in stock-out rates exceeded 40 percentage points would avert 3,320 under-five deaths and 448 over-five
for SP, DepoProvera, amoxicillin, and ACT for adults. deaths from malaria each year. In economic terms, the
Zambia is in the process of scaling up Model B aggregate household income loss averted as a result of
nationwide. A conservative estimate suggests that this the national scale-up of Model B is estimated to exceed
effort could decrease under-five mortality by 21 percent $1.6 million a year.
Source: World Bank 2010.
1. The receipt, storage, and primary distribution of drugs and other health commodities from Lusaka to approximately 120 districts
stores and hospitals are managed by a parastatal agency called Medical Stores Limited (MSL), currently under the management of
Crown Agents. Improvements of MSL have been observed in recent years as a result of large capital investments and management
reforms. Consequently, primary distribution is quite reliable and effective.
2. Zambia already had a very similar arrangement in place that has eliminated stock-outs of antiretrovirals at almost all antiretro-
viral therapy sites.

Figure 23.5 Treatment of Pregnant Women with SP/Fansidar during Antenatal Visit in Selected African Countries

80
70
60
50
Percent

40
30
20
10
0
S

IS

IS
H
H

AI

M
D
D

D
8

09
07

–0
06

08

09

08

00
00

8–
20

07
20

20

8–

20

a2
i2

00
20
00
go

bi
n

ria

na

l2
ni

m
a2
on

ha

ia

ala
ige
Be

ga
Za
an
G
ny
fC

ne
N

nz
Ke

Se
.o

Ta
ep
.R
em
D

Source: Authors’ compilation based on data from national household surveys.

Note: DHS is Demographic Health Survey. MIS is Malaria Indicator Survey. AIS is AIDS Indicator Survey.

Global Malaria Action Plan estimates at about $5 billion a
year in 2010–15 and $4.75 billion in 2020–25. The shortfall
underscores the need for policy makers and development
partners to continue to work collaboratively to ensure that
adequate resources are mobilized to scale up and sustain
control efforts over the long term.
show that with a combination of effective prevention and
treatment tools, country leadership, and partner coordina-
tion, controlling malaria in Africa is possible.
Eritrea

In 1997 and 1998 Eritrea experienced a series of malaria epi-

SUCCESS STORIES
demics that produced more than 424,000 cases; more than
This section highlights four countries that are showing sig- 500 inpatient malaria deaths were recorded in 1998 alone
nificant progress in their malaria control efforts. The profiles (RTI 2005). A 2002 Demographic Health Survey revealed

CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 425

 Box 23.4 The Roll Back Malaria Partnership

The Roll Back Malaria (RBM) Partnership is the global
framework for implementing coordinated action against
malaria. It mobilizes resources and forges consensus
among more than 500 partners, including malaria-
endemic countries, their bilateral and multilateral devel-
opment partners, the private sector, nongovernmental
and community-based organizations, foundations, and
research and academic institutions.
RBM’s strength lies in its ability to form effective part-
nerships globally and nationally. Partners work together
to scale up malaria control efforts at the country level,
coordinating their activities to avoid duplication and
fragmentation and to ensure optimal use of resources.
Source: Authors.
RBM’s overall strategy aims to reduce malaria
morbidity and mortality by reaching universal cover-
age and strengthening health systems. The Global
Malaria Action Plan defines three stages of malaria
control: scaling up for impact of preventive and ther-
apeutic interventions, sustaining control over time,
and elimination.
The RBM Partnership was launched in 1998 by
WHO, UNICEF, the United Nations Development Pro-
gramme, and the World Bank, in an effort to provide a
coordinated global response to the disease. It is served by
a secretariat, hosted by WHO in Geneva, that works to
facilitate policy coordination at the global level.

Figure 23.6 Commitments and Disbursements for Antimalaria Activities in Malaria-Endemic Countries in Africa,
2000–09

1,629
1,800
1,472
1,600

1,400

1,200 1,055
1,037
Millions of dollars

1,000

745
800 692
599
600 518
393
388
400
204
200 175
99 108
35 45 44
0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Commitments Disbursements

Source: WHO 2010b.

that less than 20 percent of households had at least one
mosquito net, just 7 percent of pregnant women slept under
a mosquito net the night before the interview, and less than
half of those used an insecticide-treated net. The report also
indicated low use of antimalaria drugs by pregnant women,
with just 5 percent of women who had given birth in the
five years preceding the survey reporting having received
antimalarial treatment for the last birth. Twelve percent of
children under five were reported as having slept under a
mosquito net the night before the interview, and just a third
of them used an insecticide-treated net; 30 percent had a
fever in the two weeks preceding the survey, but only 4 per-
cent were treated with antimalarial medication—mostly
chloroquine (NSEO and ORC Macro 2003).

426 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY

Efforts. The alarming condition in the country, coupled with
the rise of HIV/AIDS in Africa, triggered the Eritrean Ministry
of Health to establish a new strategy and program to control
HIV/AIDS, malaria, sexually transmitted diseases, and tuber-
culosis in 2000. The program, dubbed HAMSET, supple-
mented the National Malaria Control Program (NMCP),
established in 1998. The technical strategy of NMCP sought to
reduce and better target indoor residual spraying to the high-
est-risk areas and scale up rapid diagnosis and effective
treatment for fever cases, environmental management activi-
ties, and use of insecticide-treated nets. At the core of this
strategy was the strengthening of disease surveillance and
operational research activities, data from which would be used
to select and refine the mix of strategies and geographically
target control activities. During 1999–2005, the NMCP
introduced a first-line antimalarial drug and implemented
an integrated vector control program using insecticide-
treated nets, indoor residual spraying in selected areas, and
other interventions. In 2005 the Ministry of Health articu-
lated a clear strategic plan, with the goal of “reducing malaria
to such low levels that it is no longer a public health problem
in the country” (Ministry of Health 2008).
Through support from the World Bank, the Global Fund,
and other donors, Eritrea has been able to employ a com-
prehensive strategy against malaria through a centralized
approach. This approach involves targeting the use of
indoor residual spraying in the highest-risk areas, carrying
out environmental management of malaria vector breeding
sites, increasing distribution of insecticide-treated mosquito
nets, and expanding diagnostic and treatment services.
Results. The fact that about 60 percent of Eritrea’s popula-
tion of 3.5 million people live in malaria- endemic or epi-
demic-prone areas where disease is seasonal, highly focal,
and unstable puts the country in a disadvantageous position
in terms of malaria control. Despite these odds, through the
support of the World Bank and other donors, in only a few
years Eritrea has been able to introduce new first-line anti-
malarial drugs and implement an integrated vector control
program using insecticide-treated nets, indoor residual
spraying in selected areas, and other interventions. Through
these efforts, Eritrea has reduced malaria mortality by more
than 80 percent, malaria inpatient cases by 90 percent, and
malaria outpatient cases by 68 percent (WHO 2009). It also
met the target set at the Africa Summit on Roll Back Malaria
in April 2000 of reducing the incidence of malaria by at least
60 percent before 2010.
As a result of the strong commitment of the govern-
ment, Eritrea has come a long way in 10 years. Through
implementation of prevention activities, diligent environ-
mental management of malaria vector breeding sites, and
widespread diagnostic and treatment services, it has
reduced overall malaria morbidity from about 100,000 cases
in 2000 to about 8,000 in 2008 (Ministry of Health 2008)
These achievements have led to an overall decrease in
maternal and infant mortality and a healthier population.
Ethiopia
Malaria was responsible for the deaths of nearly 29,000
children—about 80 children a day—in Ethiopia in 2000
(WHO 2002). In 2003 Ethiopia had more than half a mil-
lion reported malaria cases. According to a 2005 Demo-
graphic Health Survey, malaria was the primary cause of
health problems, accounting for 17 percent of outpatient
visits, 15 percent of hospital admissions, and 29 percent of
in-patient deaths (Central Statistical Agency [Ethiopia], and
ORC Macro 2006). During this period only 5.7 percent of
households owned at least one bed net, and less than 4 per-
cent had an insecticide-treated net. The situation for chil-
dren under five years of age was dire, with only 2.3 percent
sleeping under any net the night before the survey and less
than 2 percent sleeping under an insecticide-treated net.
During the two weeks preceding the survey, about 19 per-
cent of children under five were reported as having had a
fever, and only 3 percent of these received antimalarial
drugs. Among pregnant women, only 1.6 percent slept
under any bed net the night before the survey and only
about 1 percent slept under an insecticide-treated net. Less
than 5 percent took any antimalarial drug, and only 0.5 per-
cent received any intermittent preventive treatment. Just
one-tenth of all households reported occupying a dwelling
that was ever sprayed with insecticide to prevent malaria.
Efforts. Ethiopia has been trying to control and eliminate
malaria for more than 60 years. The Malaria Eradication
Service was established in 1959, making Ethiopia one of the
pioneering countries in Africa to embark on a malaria erad-
ication effort. In 2000 the country became a cosignatory to
the Abuja Declaration, committing itself to the declaration’s
aims of increasing coverage of malaria interventions and
reducing malaria mortality by half by 2010.
The Ethiopian government has implemented a central-
ized approach to bringing the disease under control, with
significant success. In 2009, after analysis of the results of
the 2007 Malaria Indicator Survey as well as the discussions
and recommendations that followed a consultative meeting
held in Ethiopia in 2009, Ethiopia developed a six-year
CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 427
 (2010–15) National Strategic Plan for Malaria Prevention,
Control, and Elimination. In the new strategic plan, top
priorities among malaria control strategies are given to
community empowerment and social mobilization. These
priorities were based on the results of the 2007 Malaria
Indicator Survey, which showed substantial differences
between the coverage and utilization of key malaria inter-
ventions by the population at risk of malaria. Malaria diag-
nosis, case management, disease surveillance, and epidemic
control will all be geared to serve Ethiopia’s goal of shrinking
malaria-endemic areas by 2015 and eliminating the disease
throughout the country by 2020 (USAID and PMI 2010).
Donors such as the World Bank, the Global Fund, the
U.K. Department for International Development, UNICEF,
and the Canadian International Development Agency have
been providing coordinated support to the Ethiopian gov-
ernment in its fight against malaria by offering flexible and
pragmatic funding to ensure that the country’s efforts have
been scaled up and sustained. Together they have con-
tributed more than $151 million—for malaria control
efforts in the country. Partner resources have supported the
procurement and free distribution of long-lasting insectici-
dal nets, antimalarial drugs, and indoor residual spraying as
well as the strengthening of the supply chain. Other compo-
nents have included capacity building for other functions,
especially procurement and monitoring and evaluation.
Results. The effort by the Ethiopian government and its
partners has resulted in remarkable improvements in
malaria control and overall health in Ethiopia. Recent data
suggest sharp declines in the number of malaria cases as
well as declines in malaria outbreaks and deaths (Global
Fund 2009). These efforts correlated with an overall
decrease in under-five mortality by more than 20 percent
between 1990 and 2008, and maternal mortality declined
by nearly 60 percent, from 968 maternal deaths per 100,000
live births to 590.
Given the scope of the problem in Ethiopia—about 68
percent of the population of 73 million lives in areas at risk
of malaria—the government has accomplished extraordi-
nary results.
Rwanda
Rwanda has made remarkable progress in malaria control. In
2005, according to the Demographic Health Survey, only one-
fifth of households reported having at least one once-treated
mosquito net. Sixteen percent of children under five,
20 percent of pregnant women, and 13 percent of all women
428 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY
reported having slept under any mosquito net the night before
the survey. Only 5.8 percent of women reported having taken
any antimalarial drug to prevent or treat malaria during an
antenatal care visit during their last pregnancy, and just
0.3 percent of women reported having received intermittent
preventive treatment during their last antenatal care visit.
Efforts. Acting on the recommendation of its ministry of
health, the Rwandan government adopted two major strate-
gies to reduce malaria: promoting the use of insecticide-
treated nets for prevention, particularly among vulnerable
groups, including pregnant women and children under five,
and providing early diagnosis and timely treatment with
combination therapy of amodiaquine and sulfdoxine/
pyrimethamine (Concern Worldwide, International Rescue
Committee, and World Relief 2004). The prevention com-
ponent of malaria transmission was moving along steadily.
In contrast, as a result of rapid widespread resistance to
older first-line treatment antimalarial drugs, the country
was no longer able to effectively combat the disease, forcing
it to change its national treatment policy. In 2005 the
Rwandan government made ACT the official first-line anti-
malarial drug (Friends of the Global Fight against AIDS,
Tuberculosis, and Malaria 2008). To support the intense
efforts of the Rwandan government to control malaria and
strengthen health systems, development partners such as
the World Bank, the Global Fund, the President’s Malaria
Initiative, the Roll Back Malaria Partnership, WHO, Malaria
No More, the Bill & Melinda Gates Foundation, nongovern-
mental organizations, and others contributed grants and
loans to step up funding of programs across the country.
In addition to providing assistance for prevention activi-
ties such as bed net distribution and indoor residual spray-
ing and funding for antimalarial treatment, some donors
have provided grants to finance the overall strengthening of
Rwanda’s health system. In 2005 the Global Fund approved
a grant of $33.9 million to finance health systems
strengthening activities. The grant reduced the cost of health
services, improved the quality of care offered at health
centers, and financed training for more than 100 healthcare
administrators in health financing and management for
health information systems.
Results. Rwanda was once a country with extremely high
malaria prevalence and large numbers of deaths from
malaria, particularly among children. As a result of its
efforts, the number of malaria illnesses attended at health
facilities declined from 1.5 million people in 2005 to
800,000 in 2008, and the mortality rate from malaria
dropped from 41 percent in 2006 to 16 percent in 2008. In
2008 malaria was the third-leading cause of morbidity and
mortality, after pneumonia and diarrhea. The 2008 interim
Demographic Health Survey indicated a child mortality rate
of 103 deaths per 1,000 live births, a 32 percent reduction
from the 2005 rate of 152 (USAID and PMI 2009).
The percentage of households with at least one ever-
treated net increased by nearly 40 percent, from 18.1 percent
in 2005 to 57.2 percent in 2007–08. The percentage of chil-
dren under five that slept under an ever-treated net the
night before the Demographic Health Survey increased by
42 percentage points, from 15.7 percent in 2005 to 58.0 per-
cent in 2007–08, and the share of children under five with
fever in the two weeks preceding the survey decreased by
nearly 5 percentage points, from 26.2 percent in 2005 to
21.4 percent in 2007–08. By 2008 the use of health centers
had increased by 10.0 percent and the use of district hospi-
tals had increased by 16.8 percent over 2005 (Malaria Free
Future 2008).
Rwanda—a country whose entire population was once at
risk of malaria—is now one of Africa’s frontrunners in the
fight against the disease. With increasing use of bed nets,
long-lasting insecticidal-treated mosquito nets, antimalarial
drugs, and intermittent preventive treatment during preg-
nancy, Rwanda has significantly reduced the number of
malaria transmissions, malaria cases being reported at
health facilities, and deaths from malaria. These achieve-
ments have contributed to a decrease in infant and maternal
mortality and a healthier Rwanda. The Rwandan National
Malaria Control Program is continuing to pursue its very
ambitious goals, which include universal coverage of long-
lasting insecticidal-treated mosquito nets, expansion of
indoor residual spraying, subsidization of ACTs nationwide,
and better monitoring and evaluation to tackle epidemics
(Malaria Free Future 2008).
Zambia
In 2001–02 Zambia had about 3.5 million cases and 50,000
deaths a year from malaria, making the disease the leading
cause of morbidity and the second-highest cause of mortal-
ity. Disease prevalence was highest among pregnant women
and children under five, making Zambia one of the coun-
tries with the highest malaria-related maternal mortality in
Africa and one with an extremely high under-five mortality
rate (168 deaths per 1,000 live births) (DHS 2002). During
this time, only 14 percent of households in Zambia owned
an insecticide-treated net, 7 percent of children under five
slept under a mosquito net the night before the survey, and
43 percent of children under five suffered from fever or con-
vulsions in the two weeks preceding the survey, among
whom about half received antimalarial drugs for their
symptoms. Among pregnant women, just 8 percent slept
under an insecticide-treated net, and only 36 percent
received antimalarial drugs during pregnancy (DHS 2002).
Efforts. To control and reverse the deteriorating situation,
the Zambian Ministry of Health identified malaria control
as one of its main public health priorities and included it in
both the National Development Plan 2006–10 and the
National Health Strategic Plan 2006–10. Through the
National Malaria Control Centre, the Ministry of Health
developed a detailed national malaria strategic plan, which
aimed to drastically reduce the malaria burden—by scaling
up malaria control interventions—with a vision of achiev-
ing a malaria-free Zambia. This plan was the backbone for
support by the country and its partners.
The strategy adopted a two-pronged approach, aiming to
bring malaria under control with key malaria control inter-
ventions while at the same time supporting broader
improvements in health systems, including decentralizing
budgeting and planning, building capacity throughout the
supply chain for procurement and forecasting of commodi-
ties, and strengthening monitoring and evaluation and pro-
gram management. As a result of these combined efforts, 3.6
million long- lasting insecticide-treated bed nets were dis-
tributed throughout the country between 2006 and 2008,
raising the percentage of households with one long-lasting
insecticidal-treated net from 48 to 72 percent. Indoor resid-
ual spraying coverage increased from 15 districts in 2006 to
36 districts in 2008.
Results. The impact of Zambia’s efforts has been felt coun-
trywide. In just two years the number of malaria deaths
declined 47 percent, and nationwide surveys showed a
decline in parasite prevalence of 53 percent (from 21.8 per-
cent in 2006 to 10.2 percent in 2008). The percentage of
children with severe anemia declined 68 percent (from 13.3
percent in 2006 to 4.3 percent in 2008). The percentage of
children under five with malaria parasite prevalence fell by
half (from 22 percent in 2006 to 10.2 percent in 2008) and
the percentage of children under five with severe anemia fell
from 14 percent in 2006 to 4.3 percent in 2008 (MIS 2008).
These successes reduced the overall under-five mortality
from 168 in 2002 to 148 in 2008 (UNICEF 2008). The suc-
cessful scale-up of Zambia’s malaria program contributed to
a 29 percent drop in the overall mortality rate for children
under five between 2002 and 2008, saving an estimated
CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 429
 75,000 lives over that period. The capacity of the malaria
program as well as supporting critical health systems inter-
ventions was strengthened for better management.
Fragile progress. Zambia’s success illustrates one of the
most important issues associated with malaria control
efforts. Delayed external financing slowed execution of all of
the necessary malaria control interventions between June
2009 and December 2010, causing a significant decline in
malaria outcomes in some provinces of the country. Data
from the 2010 Zambia National Malaria Indicator Survey
indicates a resurgence in malaria between 2008 and 2010.
Most of this increase was in rural areas, among the poorest
segments of the population. In three of the country’s nine
provinces, about half of the decrease in malaria cases
between 2006 and 2008 was lost by 2010. In Luapula and the
Northern Provinces, there were marked declines in owner-
ship and use of mosquito nets between 2008 and 2010,
which alone could explain the increase in malaria cases.
Zambia’s experience shows that once the disease is under
control, efforts have to be sustained if gains are not to be
reversed or even worsened, because malaria can have an
even deadlier impact on a population that has lower immu-
nity to the disease.
LESSONS LEARNED AND THE WAY FORWARD
Effective tools for malaria prevention and treatment exist.
With concerted and collaborative efforts, implementation
of these tools is starting to show results in Africa. Progress
in the past few years has shown that success in Africa is pos-
sible. The results of WHO’s 2010 World Malaria Report are
very encouraging. Since 2000, 11 countries have shown
decreases in malaria cases after widescale implementation of
malaria control activities in the population at high risk. In
Botswana, Cape Verde, Eritrea, Madagascar, Rwanda,
Zambia, and other countries, the number of cases has
fallen by more than half. By the end of 2010, about 289
million insecticide-treated nets were to have been deliv-
ered to Sub-Saharan Africa, enough to cover 76 percent of
the 765 million people at risk of malaria there. Global
malaria control efforts have resulted in a reduction in the
estimated number of deaths from nearly 1 million in 2000
to 781,000 in 2009. The international community has played
an important role in supporting government efforts to
reduce the impact of malaria through substantial financial
and technical support in the past decade.
These emerging successes notwithstanding, more
remains to be done. Sustained support is needed to replace
430 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY
nets and scale up other proven interventions (access to
effective diagnosis and treatment, indoor residual spraying).
Gaps in increased coverage and use of malaria control activ-
ities remain substantial in many countries. To continue and
intensify the fight against malaria, increased funding is
required for activities that prevent and treat malaria. In
countries where there has been significant progress,
resources must be sufficient to, at the very minimum, main-
tain efforts, so that the disease does not resurface.
It is critical that political will and financing be sustained.
Failure to protect and expand on the fragile progress taking
place in Africa will lead to a resurgence of malaria cases and
deaths. Learning from early successes is critical to avoiding
such a prospect. Several lessons emerge from African coun-
tries’ experience:
■ The high level of external funding, coupled with in-
country ownership, has proven to be extremely success-
ful in combating malaria. Government ownership and
initiative to work with partners in support of established
policy in the malaria control programs at the country
level is essential.
■ Bed net coverage has rapidly increased, but other key
interventions are lagging, and the gaps are becoming
increasingly more critical. There needs to be more focus
on use of key interventions, monitoring of drug resis-
tance, deployment of rapid diagnostic tests, and access to
ACTs.
■ At a minimum, maintaining current levels of funding in
order to ensure prevention and treatment activities is
essential in the short and medium term to protect the
progress that has been achieved to date. To reach the Abuja
targets of the Roll Back Malaria Partnership—coverage of
80 percent for insecticide-treated nets for people at risk,
treatment with appropriate antimalarial drugs for people
with probable or confirmed malaria, indoor residual
spraying for households at risk, and intermittent treat-
ment in pregnancy in high-transmission areas—funding
needs to be about four times the current level. This esti-
mate does not take into account the resource needs to
reach the universal coverage target of the population at
risk by the end of 2010 set by the Secretary-General of
the United Nations and highlighted in the RBM Global
Malaria Action Plan.
■ Without continued and additional resources in the short
and medium term, countries may risk a resurgence of
malaria cases in the coming years. Zambia highlights this
issue by showing that significant progress is possible but
that it is fragile and needs a sustained effort.
■ To achieve long-term financial sustainability in national
health systems, national malaria control funding as well
as the national health budget of countries will need to
increase. In the long run, malaria control efforts will
need to be covered more substantially by those sources.
Both governments and partners need to address financial
sustainability.
■ Financing should focus on supporting countries that have
not yet scaled up their prevention and treatment efforts
while maintaining support to those that have.
■ Support needs to be provided to countries ready to move
toward eliminating malaria. In this context, more atten-
tion will have to be placed on cross-border and regional
activities.
■ Strengthening health system aspects such as supply
chains and human resources as well the interface
between health systems and disease control will be
needed to sustain gains.
NOTES
1. There are four types of human malaria: Plasmodium fal-
ciparum, Plasmodium vivax, Plasmodium malariae, and
Plasmodium ovale. Plasmodium falciparum and Plasmodium
vivax are the most common. Plasmodium falciparum, the
most deadly form of the disease, is the predominant type of
malaria in Sub-Saharan Africa.
2. The original Abuja targets were set and agreed to by 53
African heads of state in the Abuja Declaration on Roll Back
Malaria in Africa on April 25, 2000. They included the fol-
lowing targets: at least 60 percent of those suffering from
malaria have prompt access to and are able to use correct,
affordable, and appropriate treatment within eight hours of
the onset of symptoms; at least 60 percent of those at risk of
malaria, particularly pregnant women and children under
five years of age, benefit from the most suitable combination
of personal and community protective measures, such as
insecticide-treated mosquito nets and other materials to
prevent infection and suffering; at least 60 percent of all
pregnant women who are at risk of malaria, especially those
in their first pregnancies, have access to chemoprophylaxis
or presumptive intermittent treatment.
3. Lot Quality Assurance Sampling (LQAS) was developed
in the 1920s as a quality control technique for goods pro-
duced on a factory assembly line. It was used to examine a
small number of units randomly selected from each lot.
If the number of defective items in that small sample
exceeded a predetermined number, the lot was examined
more closely and either repaired or discarded; otherwise the
lot was accepted. The number of units tested and the maxi-
mum allowable number of defects were determined to ensure
that there was a high probability that the lots accepted con-
tained no more than a specified proportion of defective
goods and that the lots rejected contained a relatively high
proportion of defective goods. Since the 1980s, use of LQAS
in health assessment has increased. In 1991 WHO identi-
fied LQAS as one of the more practical rapid assessment
methods and encouraged further study. (For examples of
LQAS application in the health field, see Robertson 1997.)
LQAS has been employed for a range of purposes, includ-
ing monitoring immunization program performance in
Southeast Asia, Peru, and the United States and assessing
immunization coverage. To monitor immunization,
supervisors sample records to assess compliance with
immunization protocols. The number of “defective” or
incorrect procedures allowed per lot is set to determine
whether a facility (or lot) is accepted or rejected. LQAS
can also be used to assess compliance with a policy, such as
patient screening practices or immunization administra-
tion; identify areas of high incidence of specific diseases;
or assess program impact.
4. According to the Zambia MIS carried out in 2010, the
percentage of women sleeping under an insecticide-treated
net rose from 43 percent to 46 percent.
5. Once a diagnosis of malaria is established, effective
treatment should be started within 24 hours of the onset of
symptoms, to avoid progression to severe malaria, for which
the case fatality rate is high.
REFERENCES
Central Statistical Agency (Ethiopia), and ORC Macro.
2006. Ethiopia Demographic and Health Survey 2005.
Addis Ababa and Calverton, MD.
Concern Worldwide, International Rescue Committee, and
World Relief. 2004. Baseline Survey of the Rwanda Com-
munity Distribution of Anti-Malarials Pilot Program.
http://www.concernusa.org/media/pdf/2007/10/Rwanda
_Baseline_Malaria.
Friends of the Global Fight against AIDS, Tuberculosis and
Malaria. 2008. Country Profile: Rwanda. http://www.the
globalfight.org/view/resources/uploaded/Rwanda_Coun
try_Profile.pdf.
Gallup, J. L., and J. Sachs. 2001. “The Economic Burden of
Malaria.” American Journal of Tropical Medicine and
Hygiene 64 (Suppl. 1): 85–96.
Global Fund to Fight AIDs, Tuberculosis and Malaria. 2009.
Early Evidence of Sustainable Impact on Malaria.
http://www.theglobalfund.org/documents/publica
tions/onepagers/Malaria.pdf.
Hogan, Margaret C., Kyle J. Foreman, Mohsen Naghavi,
Stephanie Y. Ahn, Mengru Wang, Susanna M. Makela, Alan
D. Lopez, Rafael Lozano, and Christopher J. L. Murray.
CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 431
 2010. “Maternal Mortality for 181 Countries, 1980–2008: A
Systematic Analysis of Progress towards Millennium
Development Goal 5.” Lancet 375 (9726): 1609–23.
Malaria Free Future. 2008. Rwanda: Winning the Fight
against Malaria. http://www.malariafreefuture.org/
rwanda/success.php.
Ministry of Health of Rwanda. 2008. Integrated Approach to
Reducing Malaria Related Morbidity and Mortality in
Rwanda. Kigali.
NSEO (National Statistics and Evaluation Office Eritrea),
and ORC Macro. 2003. Eritrea Demographic and Health
Survey 2002. Calverton, MD.
Robertson, S. E., M. Anker, A. J. Roisin, N. Macklai, K.
Engstrom, and F. M. LaForce. 1997. “The Lot Quality Tech-
nique: A Global Review of Applications in the Assessment
of Health Services and Disease.”World Health Statistics
Quarterly 50: 199–209.
Roll Back Malaria Partnership. 2008. The Global Malaria
Action Plan—for a Malaria-Free World. http://www.rbm
.who.int/gmap/.
———. 2010. “Roll Back Malaria Report Confirms Aid for
Malaria Is Working But Predictable Long-Term Funding
Is Essential to Achieve Millennium Development Goals.”
http://www.rollbackmalaria.org/globaladvocacy/pr2010-
03-18.html.
RTI International. 2005. Combating Malaria in Eritrea.
http://www.rti.org/brochures/rti_Eritrea_malaria_broch
ure.pdf.
Statesmen’s Forum. 2010. Global Health Initiative: Integra-
tion and Innovation for Better Outcomes. June 29.
UNICEF (United Nations Children’s Fund), New York.
432 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY
UNICEF (United Nations Children’s Fund). 2008. Zambia.
http://www.unicef.org/infobycountry/zambia_statis
tics.html.
USAID (United States Agency for International Develop-
ment), and PMI (President’s Malaria Initiative). 2010a.
Country Profile Rwanda. http://www.fightingmalaria
.gov/countries/profiles/rwanda_profile.pdf
———. 2010b. Malaria Operational Plan Ethiopia FY 2010.
http://www.fightingmalaria.gov/countries/mops/fy10/
ethiopia_mop-fy10.pdf.
WHO (World Health Organization). 2007. Fact Sheet 94.
http://www.who.int/mediacentre/factsheets/fs094/en/.
———. 2009. World Malaria Report 2009. Geneva: World
Health Organization.
———. 2010a. Basic Malaria Microscopy. Part 1, Learner’s
Guide, 2nd ed. Geneva: World Health Organization.
———. 2010b. World Malaria Report 2010. Geneva: World
Health Organization.
World Bank. 2007. The World Bank Booster Program for
Malaria Control in Africa: Scaling Up for Impact. A Two-
Year Progress Report. Washington, DC. www.world-
bank.org/afr/malaria.
———. 2010. World Bank Policy Note: Enhancing Supply
Chain Management in Zambia. Africa Health Unit,
Washington, DC.
Zambia NMCC (National Malaria Control Center). 2008.
Zambia Malaria Indicator Survey 2008. Ministry of
Health, Lusaka.
———. 2010. Zambia Malaria Indicator Survey 2010. Min-
istry of Health, Lusaka.
 CHAPTER 24
Health Extension Workers in Ethiopia:
Improved Access and Coverage for the
Rural Poor
Nejmudin Kedir Bilal, Christopher H. Herbst, Feng Zhao, Agnes Soucat,
and Christophe Lemiere
ealth systems in Sub-Saharan African countries
H
often suffer from weak infrastructure, lack of
human resources, and poor supply chain manage-
ment systems. Access to health services is particularly low in
rural areas, where the majority of the population still lives.
The few private outlets that are available usually favor urban
or wealthy areas. Together with an uneven distribution of
health workers, this pattern often results in little availability
and poor quality of health services in rural areas.
Ethiopians’ access to services was particularly low before
the government came up with innovative ways of scaling
up the delivery of essential health interventions, in particu-
lar through its Health Extension Program (HEP). The HEP
was designed and implemented in recognition of the fact
that the major factor underlying poor health services in
Ethiopia is the lack of empowerment of households and
communities to promote health and prevent disease. This
chapter reviews Ethiopia’s experience in producing and
deploying health extension workers and summarizes some of
the key factors that made the program a success.1
THE PROBLEM: SHORTAGE OF HEALTH
CARE PROVIDERS IN RURAL AREAS
With a population of 80 million people, Ethiopia is the
second-most populous country in Africa. Throughout the
1990s poor nutritional status, infections, and a high fertility
rate, together with poor access to basic health services, con-
tributed to one of the highest maternal and child mortality
ratios in the world. Moreover, their use of available health
facilities was low. In 2005 almost all births took place at
home, with only 6 percent of women delivering in a clinic
or hospital (CSA and Macro International 2005). Major
causes of morbidity and mortality for children under age
five were preventable. Extrapolations from the 2005 Demo-
graphic and Health Survey showed malnutrition to be the
underlying cause of more than half of deaths of children
under five (CSA and MACRO International 2005). In 2005
only 1 percent of households owned a bed net, of which less
than 18 percent were insecticide treated.
The brunt of poor health falls on the rural poor, most
of whom live out of reach of health providers. In 2005 only
40 percent of the population lived within 10 kilometers of
a clinic or other health service delivery point. The number
of trained health workers has historically been inadequate
in Ethiopia, with shortages of almost all cadres of workers,
particularly in rural areas. Throughout the early 1990s,
universities and health professional training colleges
focused on clinically oriented training rather than on
more relevant rural-oriented community health training.
Although more than 85 percent of the population lives in
these rural areas, doctors and nurses preferred to work in
urban hospital settings, where professional opportunities
were better.
433
 ADDRESSING THE PROBLEM: THE HEALTH
EXTENSION PROGRAM
To address these problems, Ethiopia launched the HEP in
2003. The program’s objectives were to reach the poor and
deliver preventive and basic curative high-impact interven-
tions to all of the Ethiopian population. The HEP is an
ambitious government-led community health service deliv-
ery program designed to improve access to and utilization
of preventive, wellness, and basic curative services.
At the heart of this program is the production and
deployment of more than 30,000 front-line community
health workers. These health extension workers are posted
to rural communities across Ethiopia, where they provide
better and more equitable access to health services for the
poor, women, and children in a sustainable manner (Assefa
et al. 2010; Ghebreyesus 2010). The program focuses on
four major areas and provides 17 different packages to reach
the poor and address inequities (table 24.1).
Training and deployment of health extension workers
Since 2003 the HEP has been rolled out step by step to reach
full coverage of all rural villages by the end of 2010 (table
24.2; figure 24.1). As a result of the program, the ratio of
health extension workers to people increased from 1: 23,775
in 2004/05 to 1: 2,437 in 2008/09.
Health extension workers are recruited from the commu-
nities in which they will work according to specific criteria:
they are female (except in pastoralist areas), at least 18 years
old, have at least a 10th grade education, and speak the local
language. Females are selected because most of the HEP
packages relate to issues affecting mothers and children; thus
communication is thought to be easier between mothers
and female health extension workers and female workers are
thought to be more culturally acceptable. In addition, their
selection is seen as empowering women. Selection is done by
a committee made up of members nominated by the local
community and representatives from the district (woreda)
health office, the district capacity building office, and the dis-
trict education office (FMOH 2007b). Upon completion of
training, pairs of health extension workers are assigned as
salaried government employees to kebeles (neighborhoods),
where they staff health posts and work directly with individ-
ual households. Each kebele has a health post that serves
5,000 people and functions as an operational center for a
health extension worker. Five health posts and a health cen-
ter work in collaboration and for the Primary Health Care
Unit (PHCU) that serves 25,000 people. The health center
serves as a referral center and logistic hub for a health post
and also offers technical support. The health post is under
the supervision of the district health office and the kebele
administration and receives technical and practical support
from the nearby health center.
Health extension workers are trained to manage opera-
tions of health posts; conduct home visits and outreach ser-
vices to promote preventive health actions; refer cases to
health centers and follow up on referrals; identify, train, and
collaborate with voluntary community health workers; and
provide reports to district health offices.
Upon assignment, health extension workers conduct a
baseline survey of the village, using a standardized tool.
They map households and the population by age category.
They also prioritize health problems of the village, set
targets with respect to the 17 packages of services, and draft
a plan of action for the year. The draft plan of action is then
submitted to the village council and approved. The plans are

Table 24.1 Major Areas and Packages of Health Extension Program

Hygiene and environmental
sanitation
1. Proper and safe excreta
disposal system
2. Proper and safe solid and
liquid waste management
3. Water supply safety measures
4. Food hygiene and safety
measures
5. Healthy home environment
6. Arthropod and rodent control
7. Personal hygiene
Disease prevention and Health education and
control Family health services communication
1. Prevention and control of 1. Maternal and child health 1. Health education and
HIV/AIDS communication
2. Prevention and control of 2. Family planning
tuberculosis
3. Prevention and control of 3. Immunization
malaria
4. First aid 4. Adolescent reproductive health
5. Nutrition

Source: FMOH 2005.

434 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR
 Table 24.2 Number of Health Extension Workers Trained and Deployed in Ethiopia, by Region, 2005/06–2009/10
Region 2005/06 2006/07 2007/08 2008/09 2009/10 Total
Afar 0 0 164 148 196 572
Amhara 3,500 2,631 680 382 330 7,342
B. Gumuz 0 59 120 315 403 924
Dire Dawa 0 33 0 0 63 142
Gambella 0 47 0 410 0 457
Harari 0 0 0 0 8 47
Oromia 1,296 3,524 2,884 4,526 524 13,487
SNNPR 1,500 2,666 2,650 800 627 8,542
Somali 0 0 420 545 327 1,427
Tigray 840 0 0 134 73 1,442
National 7,136 8,960 6,918 7,260 2,551 34,382

Source: FMOH 2009.

Figure 24.1 Number of Health Extension Workers recognition. Health extension workers also work with
Deployed, 2004/05–2009/10 communities through traditional associations, such as idir
(community-level volunteer organizations that collect
Number of trained health

money on a regular basis to cover funeral costs and give

extension workers

34,382 some money to the family of the deceased), mehaber (pro-

24,571
17,653
9,900
2,737
2004/05 2005/06 2006/07 2007/08
Source: HSDP Annual Performance Report 2009/2010.
also disseminated to the district and regional councils and
health offices.
The basic philosophy of the HEP is to transfer ownership
of and responsibility for maintaining their own health to
individual households by transferring health knowledge
and skills to households. Health extension workers spend
75 percent of their time visiting families in their homes and
performing outreach activities in the community. The
house-to-house activity starts by identifying households to
serve as role models. These households have earned the
respect and credibility of the community because of their
extraordinary performance in other social aspects, such as
agricultural production. They are willing to change and,
upon completion of the training, are able to persuade and
convince other households to follow appropriate health
practices. The model households are considered early
adopters of health practices in line with heath extension
packages. They help diffuse health messages, leading to the
adoption of the desired practices and behaviors by the rest
of the community.
Two health extension workers are expected to train 360
model households a year. The training lasts 96 hours, after
which the household “graduates” receive a certificate as
31,831
fessional and religious organizations), ekub (a scheme in
which people regularly contribute money, which at a speci-
fied time is given to one group member; this process is
repeated until all members receive a contribution), schools,
women’s associations, and youth associations. These institu-
2008/09 2009/10
tions help communicate health messages and mobilize the
community to help with environmental cleanup, health post
construction, and other efforts.
The health extension workers’ remaining time is spent
providing services, including immunizations and injectable
contraceptives, at the health posts. They are trained to pro-
vide first aid; conduct safe and clean deliveries; diagnose
and treat malaria, diarrhea, and intestinal parasites. In 2010
the government added the diagnosis and treatment of pneu-
monia to the HEP, following an evidence-based analysis of
the potential impact of different packages of high-impact
interventions. The addition represents a significant step
toward tackling a primary cause of child mortality.
Health extension workers also participate in local politics
and are part of the multisectoral local decision-making
process. One of the two health extension workers in a village
sits on the village council, along with an elected village
chairman, a teacher, an agricultural development agent, and
a community representative. The council is the political
administration of the village that serves to prioritize the
work of the HEP, provide support to health extension work-
ers, and review their regular performance reports.
Various institutions have well-defined roles in sup-
porting the work of health extension workers once they
are deployed. The Federal Ministry of Health provides
medical equipment and supplies, such as vaccines, cold

CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR 435
 chain equipment, contraceptives, and insecticide-treated preventing HIV increasing by 78 percent in HEP villages
bednets. Regional health bureaus and district health and 46 in control villages.
offices pay the salaries of the health extension workers Vaccination coverage improved significantly. A study by
and provide technical support and political leadership. Admassiea, Abebaw, and Woldemichael (2009) finds that a
Health centers play a crucial role in providing referral care significantly larger proportion of children in villages where
and technical and practical support. In addition, volun- health extension workers were deployed were vaccinated
teer community health workers work closely with health against diphtheria, polio, and tetanus (DPT); measles; polio;
extension staff. tuberculosis; and main antigens. A study by the Center for
National Health Services indicates that more than 96 per-
cent of health posts in the three largest regions of the coun-
Impact of the intervention
try provide immunization services. A routine report from
More than 9 million households in Ethiopia (some 63 per- the Ministry of Health indicates that child vaccination is
cent of all households) have completed their training on the increasing in Ethiopia: by June 2010, 86 percent of children
17 packages of the HEP. Training and “graduation” of had received Penta 3/DPT 3 vaccine, 82 percent had
households is an output indicator for monitoring the per- received measles vaccine, and 62 percent had been fully
formance of health extension workers (table 24.3). immunized (FMOH 2010a), an average annual increase in
Health conditions, including the proportion of house- the number of fully immunized children of 15 percent
holds with access to sanitation, improved disproportion- since 2006. A household survey in the four largest regions
ately in HEP villages. A case control study conducted in finds that 64 percent of children received Penta 3 and
HEP and non–HEP villages between 2005 and 2007 indi- 68 percent of children 12–23 months had been vaccinated
cated that the proportion of households with access to against measles, one of the indicators in the Millennium
improved sanitation reached 76 percent in the intervention Development Goals (MDGs) (The Last Ten Kilometers
villages (from 39 percent at baseline). In contrast, access to Project 2009). Although the routine report and the house-
improved sanitation in the control villages increased from hold survey differ in terms of immunization coverage, both
27 percent at baseline to just 36 percent during the follow- results indicate significant improvement.
up survey period (Center for National Health Development Coverage of maternal health services also improved:
in Ethiopia 2008.) Awareness of HIV/AIDS also improved, 85 percent of health posts could provide family planning
with the level of knowledge of condoms as a means of services, 83 percent could provide antenatal care, 59 per-
cent could perform clean deliveries, and 47 percent could
provide postnatal care (Center for National Health Services
2007). Health posts had also become the major source for
Table 24.3 Number of Households “Graduating”
current users of contraceptives, as reported by about
from the HEP, as of June 2010
60 percent of women from the four largest regions; 22 per-
Number of cent of women cited heath centers. A household survey
households Number of
eligible, as of households conducted by the Last Ten Kilometers Project in the four
September “graduating,” Coverage largest regions documents that 32 percent of married
Region 2009 as of June 2010 (percent) women were practicing family planning; about 54 percent
Addis Ababa 0 n.a. 0 of these women received antenatal care services and 20 per-
Afar 258,572 130 0
Amhara 4,209,129 2,508,472 63 cent received focused antenatal care (four or more visits).
Benishangul Some 42 percent of women received at least two doses of
Gumuz 158,156 15,604 10 Tetanus Toxoid vaccine (TTII), and 54 percent were inocu-
Dire Dawa 80,041 2,400 9
Gambella 72,304 – 0
lated against tetanus. The increase in the use of any contra-
Harari 49,488 2,159 11 ceptive method among currently married women was also
Oromia 6,011,967 4,300,287 72 higher in HEP villages (where it rose from 31 percent to
SNNP 3,272,573 2,417,012 74
46 percent) than in control villages (where it rose from
Somali 708,028 30,490 17
Tigray 1,030,199 703,152 66 30 percent to 34 percent).
National 15,850,457 9,072,040 63 Ethiopia has made significant efforts in expanding cover-
Source: FMOH 2010a. age of key malaria interventions. Major scale-up efforts began
Note: n.a. = Not applicable. – = Not available. in 2004/05, with the introduction of artemisinin-based

436 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR
combination therapy (ACT) as the first line of treatment, A report published in 2009 indicates that in-patient
expanded use of rapid diagnostic tests, and enhanced vector malaria cases fell by 73 percent and deaths in children under
control and prevention through the wide distribution of five fell by 62 percent (Otten and others 2009). Another
long-lasting insecticide-treated nets coupled with targeted study reports a 48 percent reduction in morbidity, a 54 per-
indoor residual spraying. Coverage with insecticide-treated cent reduction in admission, and a 55 percent reduction in
nets has increased significantly since 2004/05 (figure 24.2). mortality related to malaria.
The 2007 Malaria Indicator Survey documents that overall In contrast to the progress made fighting malaria, post-
coverage increased by a factor of 15 between 2004/05 and natal care and assisted delivery coverage remain low. The
2009/10, with 68 percent of households in malaria-affected recently completed third round of the HEP evaluation
areas protected by at least one net or indoor residual spray- shows that less than 10 percent of women in Ethiopia
ing. Use of insecticide-treated nets by children under five receive postnatal care. Despite some improvements, many of
and pregnant women increased to nearly 45 percent in them associated with the increased availability of health
malaria-affected areas and to more than 60 percent in house- extension workers, more than 90 percent of women still give
holds that owned at least one net (FMOH 2007a). A case birth at home (Last Ten Kilometer Project 2009). The main
control study indicates that from roughly similar levels of underlying factor for the low rate of attended births is the
coverage at baseline, ownership of nets increased more in low skill levels of health extension workers in assisting deliv-
HEP villages (87 percent) than in control villages (62 per- eries and to some extent the cultural barriers to delivering at
cent) during the follow-up period (Center for National a modern health facility (FMOH 2009). To deal with the
Health Development in Ethiopia 2008). With this momen- problem, in 2009 HEP provided practical training to 5,000
tum, reaching universal coverage appears feasible. health extension workers on performing clean and safe
Residents in HEP and control villages showed a marked deliveries.
difference in seeking treatment for malaria. In HEP villages,
about 53 percent of patients with fever or malaria sought
ELEMENTS OF SUCCESS FOR HEP
treatment with antimalaria drugs the day of or the day after
the onset of symptoms. In control villages, only 20 percent Several factors contributed to the success of the HEP. This
of patients sought treatment under similar conditions. The section examines these factors.
baseline values were 33 percent and 26 percent.
Government leadership and political commitment

Figure 24.2 Number of Health Extension Workers and Ethiopia’s national health policy and strategies clearly indi-
Number of Insecticide-Treated Nets cate that accelerated training of human resources for health
Distributed in Ethiopia, 2004/05–2009/10 and expansion of infrastructure and supplies are the prior-
ity vehicles for achieving the health MDGs. The national
40,000 40 Health Sector Development Plan is a mechanism for trans-
lating Ethiopia’s health policy into action. It serves as a
35,000 35
guidebook for all stakeholders regarding priorities and
Number of workers (thousands)

30,000 30 operation modalities in the health sector.

Number of nets (millions)

Graduated HEWs Ethiopia’s health policy focuses on providing quality pro-

25,000 25
motive, preventive, and basic curative health care services in
20,000 20 an accessible and equitable manner to reach all segments of
Distributed ITNs
the population, with special attention to mothers and chil-
15,000 15
dren. The policy emphasizes establishing an effective and
10,000 10 responsive health delivery system for rural residents.
The health strategy indicates the need for an adequate
5,000 5
number of midwives and other required staff in order to
0 0 tackle the leading causes of neonatal and maternal mortality.
The Health Sector Strategic Plan III (HSDP-III) set a target
5

0
/0

/0

/0

/0

/0

/1
04

05

06

07

08

09

of training and deploying 30,000 health extension workers at

20

20

20

20

20

20

Source: FMOH 2010a. the community level and 5,000 health officers at health

CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR 437
 centers and primary hospitals to address both the preventive
and curative aspects of morbidity and mortality.
The HEP features at the top of the agenda of health
sector leaders at all levels of the health system and in their
discussions with stakeholders. In the bimonthly meeting
of the Federal Ministry of Health (FMOH) with develop-
ment partners—known as the Joint Consultative Forum
(JCF) between the Ministry of Health and development
partners—issues related to these priority areas are at the
top of the agenda. The forum also regularly reviews the
status of implementation of priorities and collectively
tackles challenges.
The other important governance structure is the
bimonthly meeting of the FMOH and the regional health
bureaus. The meeting is chaired by the minister of health
and includes the heads of the bureaus and the agencies and
departments in the Ministry of Health. The main purpose
of this steering committee is to align federal and regional
stakeholders around priorities, review progress, identify
challenges, share best practices, and agree on mitigation
measures and actions. The HEP always comes at the top of
the agenda of the steering committee. The priorities also
appear as key messages in the opening remarks made by
leaders and officials from the minister of health to district
health officers during health sector–related meetings.
The key principle underpinning the scaling up of the
HEP was ensuring ownership by the local community for
meaningful and sustainable change. Local governments at
the village and district levels, together with community rep-
resentatives, formed a committee that selects candidate
health extension workers for training based on nationally
agreed upon criteria. Local communities also contribute to
the construction of health posts, by providing construction
materials and labor. One of the two health extension work-
ers sits on the village council, which discusses the perfor-
mance and challenges facing HEP and provides support to
mitigate challenges on a regular basis. For the construction
of health centers, local governments were given the author-
ity to choose the site of construction so that it is within
walking distance to about 25,000 people.
Effective intersectoral collaboration has been demon-
strated in the last few years. Scaling up of health extension
workers and health officers would not have happened had it
not been for the strong collaboration between the Ministry
of Health and the Ministry of Education, which provided the
technical and vocational educational training (TVET) and
resources, including teachers. Strong collaboration with the
Ministry of Works and Urban Development (MWUD) and
its offices at the subnational level facilitated the construction
438 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR
of more than 2,500 health centers. MWUD’s support in
providing contractors, supervision, and quality assurance of
construction has been one of the key critical success factors
in this area.
The most important success factor has been the increase
in expenditure allocated to implement the HEP (figure
24.3). Total per capita health expenditure in Ethiopia was
$7.13 per capita in 2005 (FMOH 2005), 28 percent of it
borne by the government. The MDG needs assessment con-
ducted in 2004 recommended an additional $3.48 per capita
per year on average over the period 2005–15 from all
sources, peaking at $4.55 per capita in 2015 just for scaling
up the HEP. To meet this guideline, the government com-
mitted to at least double its share of health expenditure over
the life of HSDP-III (2005–10). By 2008 total health expen-
diture had reached $16.10 per capita, following an increase
in government expenditure to health of 77 percent over
three years, indicating that the government is on track to
fulfill its commitment over five years (FMOH/Abt 2010).
In addition to increasing expenditure, the federal gov-
ernment negotiated with the regional governments to share
the cost of implementing the national health strategy. Roles
and responsibilities were clarified and agreed upon by the
FMOH and the subnational health authorities (FMOH
2009). Local governments (regions, zones, and districts)
took responsibility for covering the full cost of constructing
health posts, partially covering the cost of construction of
health centers ($230 million), and fully covering the salaries
of health extension workers and health officers ($143 mil-
lion) over five years. The government supports these activi-
ties, through a commitment of Br 275 million (about $21
million) a year for the salaries of health extension workers.
The agreement defining the roles of different levels of the
health system in scaling up health interventions is signed by
the minister of health and the heads of the 11 regional
Figure 24.3 Total Health Expenditure in Ethiopia,
1995/96–2007/08
20
16.1
15
dollars
10
7.13
5.6
5 4.09
0
1995/96 1999/2000 2004/05 2007/08
Source: FMOH/Abt Associates 2010.
health bureaus each year at the annual review conference. Figure 24.4 Cumulative Number of Health Posts
This system set a model for using development partner Constructed in Ethiopia, EFY 2004–09
funds as a catalytic fund to provide incentives to different
levels of health authorities to allocate more resources to No. of health posts
16,000
agreed-upon priorities in the health sector.
14,000 14,192
12,000 12,488
Multifaceted, systemwide approach 10,000
10,621

Number
A multipronged strategy that would address systemic weak- 8,000 8,528
nesses was at the heart of the development of the HEP. Sim- 6,000 6,191
ply increasing the number of health extension workers was 4,000
not viewed as likely to be effective in isolation. The increase
2,000
had to be implemented together with improvements in
0
other system requirements.

0
/0

/0

/0

/0

/1
One important factor in the success of the HEP in rural

04

06

07

08

09
20

20

20

20

20
areas was the strategic commitment of regions and districts
Source: FMOH 2009.
to address complementary and supportive factors, including
the availability of other health cadres; the availability of
health infrastructure (health posts and health centers) Table 24.4 Ratio of Health Human Resources in
together with adequate supplies of equipment and pharma- Public Sector to People in Ethiopia,
ceuticals; and health information systems. All of these 2001/02–2008/09
strategic elements have been addressed, with various
Category 2001/2002 2004/2005 2008/2009
degrees of success.
All physicians 1:35,603 1:35,604 1: 34,986
General practitioners 1: 54,385 1:58,203 1:76,302
Infrastructure. One of the components of the sector Health officers 1:138,884 1:104,050 1: 20,638
Nurses (excluding
strategy was the construction and rehabilitation of health
midwives), BSc and
facilities. The number of health posts staffed by health diploma 1:5,613 1: 4,980 1: 4,895
extension workers increased from 6,191 in 2004 to 14,192 Midwives (senior) 1:77,981 1: 55,782 1: 57,354
in 2010 (FMOH 2009) (figure 24.4). Funds were secured Health extension n.a. 1: 23,775 1: 2,437
workers
from development partners to equip all health posts. As of
2009, health posts were present in 70 percent of the kebe- Source: FMOH.
les of Ethiopia’s four largest regions (Last Ten Kilometer Note: n.a. = Not applicable.

Project 2009).
As of 2007 about 94 percent of villages had health posts
built by the local communities or the government (Center
for National Health Development in Ethiopia 2007). The
majority of these health posts were equipped with basic fur-
niture. However, access to other important services (clean
water, electricity, telephone, and means of transportation)
was generally low in all regions.
Medical cadres for supervision, support, and refer-
ral of health extension workers. The training of
health extension workers was part of a larger government
effort to increase the number, distribution, and perfor-
mance of health workers. Over the past few years, there has
been a significant increase in the intake and output of
health training colleges and universities for other categories
of health workers (table 24.4). The output of nurses has
increased, boosted by the proliferation of private training
institutes. The number of universities involved in training
of medical doctors also increased, from three to six, over a
decade, raising the average annual output of doctors from
90 to 350. One of the most significant achievements of
these universities was the production of cadres with inter-
mediate skill mixes adapted to the Ethiopian context.
In 2005 the government began preparing a new midlevel
health cadre called health officers. Health officers provide
most of the curative care services at the first level of clinical
services (health center) level. The program was launched
with the financial support of the U.S. Agency for Interna-
tional Development and the technical support of the Carter
Center. A network of 5 universities (Jimma, Haramaya,
Hawassa, Mekelle, and Gondar) and 21 hospitals in

CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR 439
 7 regions was strengthened to provide theoretical and prac-
tical training to trainees. The accelerated training of health
officers lasts five years, including three years at university
and two years in training hospitals. After completion of
their training, health officers are supposed to provide sup-
port to the HEP by working in health centers with other
health team members in providing curative, preventive,
promotive, and rehabilitative health care. Their duties
include the following:
■ Assess community health needs
■ Plan, implement, and evaluate activities and resources of
the primary health care unit
■ Collect, organize, and analyze health and health-related
data from health institutions, communities, and other
relevant areas and use and disseminate the information
to the community and other concerned bodies
■ Conduct and provide on-the-job training to the staff of
the primary health care unit and to community health
workers
■ Provide comprehensive outpatient care and in-patient
services
■ Perform minor surgical procedures
■ Refer difficult cases and follow up on return to ensure
continuity of care
■ Mobilize individuals, families, and communities for
health action
■ Promote and be engaged in intersectoral activities
■ Undertake essential and operational health research
■ Document and report all primary health care unit
activities
■ Oversee the equity and efficiency of health services
Public health officers can go on to become physicians or
public health specialists, or they can pursue master’s level
training on emergency surgery.
The health officers program has more than tripled the
uptake and production of health officers training in
Ethiopia (see table 24.4). More than 5,500 health officer stu-
dents have entered the training program. By August 2010,
3,871 (71 percent) of the overall target had graduated.
An independent study of graduating students (Carter
Center 2009) reveals the following:
■ 90–97 percent feel prepared to diagnose and manage or
refer common adulthood infectious diseases and com-
mon chronic illnesses.
■ 80–90 percent feel prepared to diagnose, manage, or refer
surgical emergencies (acute abdominal problems) and
440 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR
perform minor surgical procedures (wound and trauma
management).
■ 80–95 percent feel confident to provide antenatal and
postnatal care and to diagnose and manage or refer nor-
mal labor and common gynecologic problems.
■ 55–65 percent feel confident to perform common obstet-
ric and gynecologic procedures, such as manual removal
of placenta, instrumental delivery (vacuum), safe abor-
tion, and postabortion care. The low level of confidence
is attributable mainly to inadequate on-the-job training
caused by a shortage of obstetricians, surgeons, and sur-
gical equipment and supplies.
Accountability structure for health extension
workers. A supportive accountability mechanism was
established to support health extension workers. Supervi-
sors were trained and deployed in 3,200 health centers. Each
supervisor supports 10 health extension workers in 5 satel-
lite health posts, which together form a primary health care
unit. An independent survey indicates that two-thirds of the
health posts were supervised during the three months pre-
ceding the survey and that 75 percent of these health posts
received feedback from supervisors (Carter Center 2009).
Adequate supplies and equipment. Ensuring continuous
logistics supply—equipment, contraceptives, vaccines,
insecticide-treated nets, delivery kits, and so forth—is a cru-
cial area of support to health extension workers. At the fed-
eral level, FMOH engaged in a continuous dialogue with
development partners to improve aid effectiveness by sup-
porting priorities (such as the HEP) and improving harmo-
nization and alignment to reduce transaction costs. Two
pooled fund arrangements, Protection of Basic Services
and the MDG Performance Fund, were established and
made functional at the federal level. The establishment of
these funds contributed significantly to financing the pro-
curement of contraceptives, vaccines, insecticide-treated
nets, and equipment. In addition, funds from global health
initiatives, particularly the Global Fund to Fight AIDS,
Tuberculosis and Malaria (GFATM) and the Global Alliance
for Vaccines and Immunisation (GAVI), contributed to the
financing and procurement of equipment and pharmaceu-
tical supplies as part of their support to health system
strengthening.
Ensuring timely and continuous logistic supply requires
an efficient procurement and distribution system. Building
such a system takes time. Ethiopia had to come up with a
quick transitional mechanism while building sustainable
capacity of the health sector in supply chain management
through the implementation of the logistic master plan.
One mechanism it adopted to do so was the use of develop-
ment partners with comparative advantage in procurement
and distribution of commodities to the health posts. FMOH
negotiated with UNICEF and the United Nations Popula-
tion Fund (UNFPA) for the procurement and distribution
of vaccines, cold chain equipment, insecticide-treated
nets, contraceptives, and health post kits. This partnership
provided a remarkable short-term solution to the logistic
challenges associated with the HEP.
Information systems. Information systems that facilitate
the collection, analysis, use, and dissemination of data
were perceived as significantly improving the support pro-
vided to the HEP as well as the quality and relevance of the
HEP to beneficiary communities. Accordingly, the FMOH
designed a robust, simplified, and standardized health
management information system contextualized to the
Ethiopian setting. Family folders were developed based on
the 17 packages of health interventions, and health exten-
sion workers and HEP supervisors were trained on the sys-
tem’s application and use. Each household has a family
folder that records the status of its members (for family
planning, antenatal care, expanded program of immuniza-
tion, and so forth) and the household in general (ownership
and use of a latrine, clean water supply and use, waste dis-
posal, and so forth) in terms of completing the desired
changes indicated in the HEP. The Ministry of Health is
printing family folders to make sure that all households in
Ethiopia have a formal medical record.
Innovative training strategy
Training more than 30,000 health extension workers could
not have been done through traditional means. Innovative
approaches were applied through the use of existing TVET
for theoretical training and health centers for practical
training. FMOH provided training materials; regional
health bureaus provided the stipend and transportation ser-
vices for the students.
Health extension workers must complete a 12-month
course of theoretical and field training. One-quarter of the
period is allocated to theoretical teaching at TVET institu-
tions; three-quarters of the period is spent in a practicum in
the community. The average monthly cost of the program
per health extension worker was about $234 for training,
$178 for apprenticeships, and $83 for salaries.
The TVET institutions are run by the Ministry of Educa-
tion. The number of institutions involved in the training of
CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR
health extension workers grew every year, reaching 36 by
2007. The same year, 140 TVET tutors were given trainers
instruction in HEP teaching methodology as well as an inte-
grated refresher training course. Appropriate education
materials were developed, printed, and distributed to the
TVET institutions.
Mobilizing financial support from development
partners
The progressive increase in domestic resource allocation to
priorities was key to ensuring sustainability. With regard to
the HEP, an agreement was reached between FMOH and
regional health bureaus under which the ministry mobilizes
funds from development partners to provide support to the
TVET institutions for printing the HEP training manuals
and tools and for procuring and distributing medical equip-
ment, insecticide-treated nets, contraceptives, and other
supplies; subnational governments allocate domestic
resources for stipends to health extension workers during
training, pay their full salary on deployment, and cover the
costs of building the health posts.
FMOH mobilized resources from development part-
ners to provide teaching materials, transportation ser-
vices, and other relevant supplies for the accelerated
training of health officers. To put in place 3,200 health
centers by the end of 2010, the regional health bureaus
agreed that subnational authorities would construct one
matching health center for every health center the
FMOH constructed using donor funds. As part of this
agreement, FMOH covers the required medical equip-
ment while the regional health bureaus cover the cost of
furnishing the centers.
Significant support was mobilized from various devel-
opment partners to support the new initiatives. GFATM
and GAVI provided financial support for the construction
of 512 health centers and procurement of medical equip-
ment for 7,340 health posts and 300 health centers (table
24.5). Development partners that contributed to the Pro-
tection of Basic Services program—the World Bank, the
United Kingdom’s Department for International Develop-
ment (DFID)], RNE, Irish Aid, Canadian International
Development Association (CIDA) , Italian Cooperation—
also provided support for the procurement of medical
equipment for 2,295 health centers and 7,000 health posts,
contraceptives worth $37 million, essential drugs worth
$10 million, and insecticide-treated nets worth $7.9 mil-
lion. These contributions helped provide the required
inputs to maintain HEP basic services.
441

Table 24.5 Support from Development Partners in
Key Areas of Scaling Up of Health
Interventions, 2006–2010
Support
Development partner/area (millions of dollars)
Global Alliance for Vaccines and Immunisation (GAVI)
Health center construction (212)
Health center equipment (300)
Health post equipment (7,340)
Subtotal
Global Find to Fight AIDS, Tuberculosis and Malaria (GFATM)
Health center construction (300)
Protection of Basic Services I and II (World Bank, DFID,
RNE, Irish Aid, Canadian CIDA, Italian Cooperation)
Health center equipment (1,999)
Health post equipment (7,000)
Contraceptives
Insecticide-treated nets
Vaccines and cold chain equipment
Artemisinin-based combination
therapy (ACT)
Essential drugs
Subtotal 142.7
MDG Performance Fund
Health center construction
Health management Information
system printing
Contraceptives
Vaccines
Advocacy and training
Insecticide-treated nets
Subtotal
USAID
Accelerated training of health officer
(5 universities and 21 teaching
hospitals)
Total 287.4
Source: Authors’ compilation, based on data from FMOH and
Financial Resource Mobilization Directorate.
In addition to their own financial support, UNICEF and
UNFPA helped in the bulk procurement and distribution of
health commodities to the service delivery points by using
the funds provided by the Global Fund, GAVI, the World
Bank protection of basic services (PBS), and the MDG Per-
formance Fund. UNICEF provided support in procuring
and distributing insecticide-treated nets, vaccines, and
health post kits to health facilities. UNFPA supported the
government by procuring and distributing contraceptives.
LESSONS LEARNED
The HEP has been successful largely because of the strong
political commitment to strengthening health systems, with
the ultimate goal of improving coverage of and access to
442 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR
health services by the rural poor. Investment in health
extension workers has been part of a wider package of sup-
port services that is showing promising results.
Some weaknesses remain, and not all regions show a sim-
ilar level of success. Incentive packages for health human
resources are lacking, and the link between compensation
25.8 and performance is weak. The capacity of the district health
6.8 offices to provide supervision, monitoring, and evaluation is
20
low. These and other constraints need to be addressed if gaps
52.6
in coverage are to be filled and access scaled up even further.
55.4 A number of lessons can be drawn from the imple-
mentation of the HEP, which can be replicated in other
31.2 countries:
5.6
43.6 First, political leadership and champions at higher levels
31.6 are a critical success factor in improving health out-
9.8
comes. Adoption of the HEP as a major political agenda
2.4 of the government at various levels of the health system
18.5 tightened the focus and galvanized the involvement of
various stakeholders. Beyond the general increase in its
fiscal space to finance the HEP, the government made
15
sure that such increments happened at local levels.
6.5 Accordingly, salaries of health extension workers, con-
3.1 struction of health posts, and the basic running cost of
5 the HEP are financed mainly by subnational health
2.3 authorities (regions, districts, and zones), creating the
4.8
36.7
foundation for local ownership and sustainability of the
program.
Second, delivery of services and management of programs
should be integrated into existing systems. Vertical
11.8 programs and projects can be successful in the short
term, but they are often unsustainable. What HEP has
demonstrated is that vertically mobilized resources can
be used for systemwide interventions that make dis-
ease-specific programs successful while strengthening
health systems. Adopting this approach avoids creating
parallel systems and procedures in the delivery of ser-
vices and management of programs, averting unneces-
sary administrative burdens, transaction costs, and
inefficiencies.
Third, community ownership is a key to sustainable
impact. The major principle underpinning the HEP is
transferring the right knowledge and skills to communi-
ties and households so that they are able to adopt behav-
iors that improve their own health. Households are
trained and certified, after which they take responsibil-
ity for promoting behaviors that lead to positive health
outcomes.
Fourth, all components of the health system need to be
addressed to make a program work. The HEP does not
merely train and deploy health extension workers. Sig-
nificant investment is made in setting up and equipping
 health posts to serve as formal institutional hubs for the
program. A health information system has been adapted,
and HEP supervisors have been trained and deployed to
enhance supportive supervision and continuous
improvement in quality of program management and
service delivery. Continuous assessment and in-service
training have been conducted to fill the gap in capacity of
health extension workers. Referral levels (to health cen-
ters, health officers, and so forth) have been expanded to
ensure delivery of a complete package of essential ser-
vices. A major shift in the amount and modality of
financing of the sector has been undertaken to support
the key components of the program.
Fifth, buy-in and involvement of key stakeholders is crucial.
A unique model for partnership and collaboration has
evolved in Ethiopia between the government and various
actors in the health system, including the community,
development partners, and other sectors. The growing
trust among these stakeholders has resulted in harmo-
nization of financing, program implementation, moni-
toring, and evaluation, leading to further strengthening
of health systems.
Sixth, a program needs to be flexible and adaptable to vari-
ous contexts. The HEP has been implemented in settings
with significant diversity in socioeconomic, cultural, and
geographic conditions without compromising the basic
principles that led to its success. Ethiopia designed three
versions of the HEP (agrarian, urban, and pastoralist) to
modify and fit key aspects of program implementation
in these widely varying contexts. This flexible nature of
the program provides key lessons that are unique to the
different environments, facilitating replication in other
countries.
Seventh, success of a program or an intervention should be
assessed by concrete and measurable improvements in
health outcomes. Implementation of HEP has shown
encouraging results in a short time in increasing cover-
age of essential interventions and reducing morbidity
and mortality related to communicable health diseases.
NOTE
1. Certain criteria were applied to identify an intervention
as successful. An intervention is considered successful if it
addresses diseases of public health importance; is owned
and financed by the government (to ensure sustainability);
fits into the country’s conventional health system without
creating parallel structures and systems; is flexible enough
to be applied in different socioeconomic, cultural, and geo-
graphic settings; is embraced and supported by develop-
ment partners, nongovernmental organizations, and other
stakeholders; is delivered at low cost; and shows concrete
results in terms of improving health outcomes.
CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR
BIBLIOGRAPHY
Admassiea, A., D. Abebawa, and A. Woldemichael. 2009.
“Impact Evaluation of the Ethiopian Health Services
Extension Program.” Journal of Development Effectiveness
1 (4): 430–49.
Assefa, A., Alebachew, H. Fassil, N. Haniko, M. Ihalainen,
A. Zwandor, and B. Kinfegebriel. 2010. “HIV/AIDS and
the Health-Related Millennium Development Goals. The
Experience in Ethiopia.” Lancet.
Carter Center. 2009. Evaluation of the Quality of Training of
Health Officers and Quality and Utilization of Health
Learning Materials. Report of an Independent Assess-
ment. Addis Ababa.
Center for National Health Development in Ethiopia. 2008.
Ethiopia Health Extension Program Evaluation Study,
2005–2007, vol. I. Household Health Survey. Addis Ababa.
Center for National Health Services, 2007. Functioning of
Health Posts. Addis Ababa.
CSA (Central Statistical Authority), and MACRO Interna-
tional. Ethiopia Demographic and Health Survey, 2005.
Addis Ababa.
FMOH (Federal Ministry of Health). 2005. Third National
Health Account Report. Addis Ababa.
———. 2007a. Ethiopian National Malaria Indicator Sur-
vey. Addis Ababa
———. 2007b. Health Extension Program in Ethiopia Pro-
file. Addis Ababa.
———. 2009. Report of the Joint Review Mission of HSDP
III. Addis Ababa.
———. 2010a. Annual Performance Report of HSDP. Addis
Ababa.
———. 2010b. Ethiopian National Malaria Indicator Sur-
vey. Addis Ababa.
FMOH (Federal Ministry of Health), and Abt Associates.
2010. Fourth National Health Account Report. Addis
Ababa.
Ghebreyesus, A. T. 2010. “Achieving the Health MDGs:
Country Ownership in Four Steps.” Lancet 376 (9747):
1127–28. http://www.lancet.com/journals/lancet/article/
PIIS0140-6736(10)61465-1/fulltext?_eventId=login.
Last Ten Kilometers Project. 2009. Baseline Household
Health Survey: Amhara, Oromiya, SNNP and Tigray. JSI
Research & Training, Inc., Addis Ababa.
Otten, Mac, Maru Aregawi, Wilson Were, Corine Karema,
Ambachew Medin, Worku Bekele, Daddi Jima, Khoti
Gausi, Ryuichi Komatsu, Eline Korenromp, Daniel Low-
Beer, and Mark Grabowsky. 2009. “Initial Evidence of
Reduction of Malaria Cases and Deaths in Rwanda and
Ethiopia Due to Rapid Scale-Up of Malaria Prevention
and Treatment.” Malaria Journal 8 (14).
443
 CHAPTER 25
Family Planning Trends in Sub-Saharan
Africa: Progress, Prospects, and
Lessons Learned
Mona Sharan, Saifuddin Ahmed, John May, and Agnes Soucat
ub-Saharan Africa has the highest average fertility
S
rate in the world. In 2009 the total fertility rate
(TFR), or the average number of births per woman,
was 5.1—more than twice that in South Asia (2.8) or Latin
America and the Caribbean (2.2) (World Bank 2009). The
average contraceptive prevalence rate (22 percent) is less
than half that of South Asia (53 percent) and less than a
third that of East Asia (77 percent) (World Bank 2009).
Population in the region continues to grow at a faster rate
(2.3 percent) than in other regions, including both Asia and
Latin America (1.1 percent each) (UN DESA 2008).
These dismal indicators at the aggregate level conceal
ongoing and imminent fertility transitions that are taking
place at the country level. Contrary to the popular percep-
tion, there is evidence of progress in fertility decline in many
countries across Africa (Cohen 1998). Emerging research
shows that fertility transition has already begun in Sub-
Saharan Africa and that some countries are undergoing
dynamic and unprecedented changes in fertility patterns.
Acceptance of family planning in the region has tradi-
tionally been low and cultural resistance to family planning
high (Caldwell and Caldwell 1987). Nevertheless, over the
past two decades, contraceptive use increased in several
countries. Its impact, along with that of changes in other
determinants of fertility, is leading the onset of fertility
decline in the region.
This chapter examines contraceptive use and fertility
trends in Sub-Saharan African countries between 1986
and 2009. It is based on an analysis of household survey
data from Demographic and Health Surveys in 45 coun-
tries (all countries with more than one survey during the
reference period were included). Data from the World
Bank’s World Development Indicators and the United
Nations were also used for additional analyses. The chapter
discusses the impact of micro and macro level factors,
that have contributed to changes in the dynamics of con-
traceptive use and fertility decline. It highlights national
policies, institutional frameworks, and service delivery
strategies in selected countries in which the greatest
progress has been made.
TRENDS IN CONTRACEPTIVE USE
Contraceptive use indicators—including the contraceptive
prevalence rate (CPR), method mix, and unmet need—are
showing encouraging progress in some countries of the
region. Such changes are often precursors of fertility decline.
Changes in the contraceptive prevalence rate
The modern contraceptive prevalence rate—the proportion
of women of reproductive age who are using a modern con-
traceptive method—varies widely across the region. Among
women of reproductive age, CPRs for modern methods
ranged from 1.2 percent in Somalia to 60.3 percent in South
Africa (figure 25.1). Geographic variations in family plan-
ning use were apparent in the findings, with countries in
Southern Africa reporting the highest levels of contraceptive
445
 Figure 25.1 Modern Contraceptive Prevalence Rates in Sub-Saharan Africa, by Country

Somalia 1.2
Chad 1.7
Guinea 4
Angola 4.5
Niger 5
Eritrea 5.1
Congo, Dem. Rep. of 5.8
Benin 5.9
Sierra Leone 6
Guinea-Bissau 6.1
Mali 6.3
Côte d’Ivoire 8
Mauritania 8
Burundi 8.5
Central African Republic 8.6
Nigeria 9.1
Senegal 10
Liberia 10.3
Togo 11.1
Gabon 11.8
Mozambique 11.8
Cameroon 12
Congo, Rep. 12.7
Gambia, The 12.7
Burkina Faso 13.3
Ethiopia 13.7
Ghana 16.6
Madagascar 16.7
Djibouti 17.1
Uganda 17.9
Comoros 19.3
Rwanda 26.1
Zambia 26.5
São Tomé and Princípe 27.4
Kenya 31.5
Lesotho 35.2
Malawi 38.4
Mauritius 39.3
Botswana 42.1
Swaziland 46.8
Namibia 53.5
Zimbabwe 57.9
South Africa 60.3
0 10 20 30 40 50 60
Percent
East Africa Central Africa West Africa Southern Africa

Source: United Nations Population Division 2009.

use, followed by countries in East Africa. With a few excep-
tions, West and Central African countries report very low
rates of family planning use. Some of the lowest contracep-
tive prevalence rates in the world exist in these two subre-
gions of Africa.
Levels of fertility were high throughout Sub-Saharan
Africa during the 1970s. Some indication of fertility transi-
tion began to emerge in the 1980s in some parts of Africa.
Evidence was accumulating that fertility was falling or
expected to fall and contraceptive prevalence was high in

446 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
Southern Africa, at least in comparison with the levels for
Africa as a whole in the 1970s (Lucas 1992). Parts of East
Africa were also showing signs of change. Surveys carried
out in the mid-1980s reported pockets of high contraceptive
prevalence in Botswana, Kenya, and Zimbabwe and an
increasing desire among more than one-third of women in
these countries to stop childbearing—a proportion that was
above the 10 percent average for the region (Way, Cross, and
Kumar 1987).
More recent data corroborate the onset of fertility decline
in parts of the region. Wide variation in contraceptive preva-
lence persists across countries, suggesting that country-level
contexts and policies may underlie these differentials.
Although overall progress is only modest, the experience
of a few countries in increasing contraceptive prevalence
stands out. An analysis of fertility trends in 23 countries of
Sub-Saharan Africa from 1980 to 1995 showed evidence of
fertility decline in two-thirds of the countries, with a par-
ticularly rapid decline in Kenya and Zimbabwe (Kirk and
Pillet 1998).
Trend data on modern CPR over the past 20 years show
that some countries have made remarkable progress
(figure 25.2). Countries such as Namibia and Zimbabwe
started out with high levels of contraceptive prevalence in
the 1990s and saw their rates climb steeply over the next
two decades. Other countries, such as Malawi, Madagas-
car, and Mozambique, began with relatively lower CPRs in
the early 1990s, but these rates sharply increased in the
following years. Progress was apparent not only in South-
ern Africa but also in countries in East Africa, where
increases in Zambia, Uganda, and Tanzania were particu-
larly noteworthy.
CPRs between the first and the most recent Demo-
graphic and Health Survey during the study period were
compared to examine the rate of change over time. The
findings reveal that some countries experienced dramatic
increases in contraceptive prevalence within relatively
short periods of time. The increases in CPR in Namibia,
Tanzania, Zambia, and Zimbabwe were particularly rapid.
Mozambique experienced the steepest increase in modern
CPR within the shortest time frame in the region:
between 1997 and 2003 its CPR increased more than four-
fold, from 5.6 to 25.5 percent. Although other countries
reported larger increases, changes there occurred over
longer time periods. Data show that between the first and
the most recent Demographic and Health Survey, all
countries reported increases in CPR. Malawi, Mozam-
bique, Namibia, Zambia, and Zimbabwe made the great-
est progress.
CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
Analysis comparing African countries with countries in
other regions shows that the rate of percentage change in
CPRs in Malawi, Mozambique, Namibia, and Zambia was far
greater than in many South Asian, East Asian, and Latin
American countries. Although fertility transition began late
in Africa, the increase in CPR occurred from a very low base,
thus there was scope for a greater magnitude of change rela-
tive to countries in other regions that had already attained
high levels of coverage (figure 25.3).
Changes in the choice of contraceptive method
An indicator of progress in family planning adoption is the
change in the type of contraceptive methods used by family
planning acceptors. The use of traditional methods tends to
be higher in settings where acceptance of family planning is
low and use of family planning programs is weak. Tradi-
tional methods have a high failure rate compared with mod-
ern methods and are therefore not considered an effective
mode of contraception.
Trends in contraceptive choice show that in many
countries of the region, use of traditional methods has
declined and use of modern methods increased
(figure 25.4). The use of modern methods has increased
most markedly in countries that had the greatest increases
in CPR (Madagascar, Malawi, Namibia, Zambia, and Zim-
babwe). Use of traditional methods in these countries has
either remained stagnant or has decreased. Ghana, Kenya,
Tanzania, and Uganda showed increases in use of modern
methods while maintaining use of traditional methods. In
West African countries such as Benin, Burkina Faso,
Cameroon, Senegal, and Togo, traditional method use
declined and relatively modest gains in modern method
use were observed.
Family planning programs that have been successful in
Africa have promoted birth spacing. Marriage patterns in
Africa differ from those in Asia, possibly accounting for a
cultural preference for spacing methods. Various studies
in the region document African cultural preferences for
spacing rather than limiting births (Cohen 1998). In con-
trast to Asian family planning programs, which have
emphasized permanent contraceptive methods, such as
sterilization and abortion, programs in Africa rely on
temporary methods, such as pills, injectables, and
implants (Caldwell and Caldwell 1988). It has been sug-
gested that successful program strategies in Africa must
promote methods that are temporary, can be used
covertly by women, and do not have to be stored at home
(Caldwell and Caldwell 2002).
447
448

Figure 25.2 Trends in Modern Contraceptive Prevalence Rates in Selected Countries in Sub-Saharan Africa
Angola Benin Botswana Burkina Faso Burundi Cameroon Central African Republic Chad
60

40

20

Comoros Côte d’Ivoire Congo, Dem. Rep. of Djibouti Eritrea Ethiopia Gambia Ghana
60

40

20

Guinea Guinea-Bissau Kenya Lesotho Liberia Madagascar Malawi Mali

60

40

20

Mauritania Mauritius Mozambique Namibia Niger Nigeria Rwanda Senegal

60

40

20

Sierra Leone Somalia South Africa Swaziland Togo Uganda Zambia Zimbabwe
60

40

20

0
80

90

00

10

80

90

00

10

80

90

00

10

80

90

00

10

80

90

00

10

80

90

00

10

80

90

00

10

80

90

00

10
19

19

20

20

19

19

20

20

19

19

20

20

19

19

20

20

19

19

20

20

19

19

20

20

19

19

20

20

19

19

20

20
East Africa Central Africa West Africa Southern Africa

Source: Demographic and Health Surveys, various years.

449

Percentage change

0
1
2
3
C
ha
In d
d
N ia
M ige
or r
Vi occ
et o
n
Li am
b
Se eria
ne
ga
M l
Health Surveys

Er ali
i
G trea
ui
ne
Be a
N ni
Az ige n
er ria
ba
ij

Source: Demographic and Health Surveys, various years.

Ph T an
ilip og
pi o
n
G es
R ha
Bu w na
rk an
in da
aF
a
G Yemso
ua e
te n
G ma
H eor la

Sub-Saharan Africa
o g
El ndu ia
Sa ra
Zi lva s
m do
ba r

Asia
bw
Ke e
n
C Uga ya
am nd
Ba er a
ng oo
la n
Pa des
kis h
t
Jo an
rd
a
H n
N iti a
e
Bo pal
In liv
do ia
North Africa/Central Asia

n
U esia
Takrai
C nza ne
ol ni
o a
Ec mbi
N uad a
ica o
r
C E ragu
ô
Latin America

D te thi a
om d op
in C ’Ivo ia
ica am ir
n bo e
Re d
pu ia
Eg
yp M Za blic
t, ad mb
Ar ag ia
ab as
Re car
p.
o
Pe f
M ru
ala
B wi
Pa raz
Figure 25.3 Annual Rate of Change in Modern Contraceptive Use in Selected Countries between First and Last Rounds of Demographic and

ra il
g
M Na uay
oz m
am ibi
bi a
qu
e
450

Figure 25.4 Trends in Modern and Traditional Contraceptive Prevalence Rates in Selected Countries in Sub-Saharan Africa

Benin Burkina Faso Cameroon Chad Côte d’Ivoire

50

25

Eritrea Ethiopia Ghana Guinea Kenya

50

25

Liberia Madagascar Malawi Mali Mozambique

50
Percent

25

Namibia Niger Nigeria Rwanda Senegal

50

25

Tanzania Togo Uganda Zambia Zimbabwe

50

25

0
80

90

00

10

80

90

00

10

80

90

00

10

80

90

00

10

80

90

00

10
19

19

20

20

19

19

20

20

19

19

20

20

19

19

20

20

19

19

20

20
modern methods traditional methods

Source: Demographic and Health Surveys, various years.

 The use of family planning methods depends not just on
users’ preferences but also on health system characteristics.
Strong family planning programs rely on effective family
planning service delivery strategies, such as those that
offer methods tailored to the needs of users, provide family
planning counseling and medical expertise for administer-
ing methods, and follow up on users’ response to the
method. Countries in the region with frail health systems
are faced with the challenge of improving contraceptive
method choice within existing constraints. There is growing
evidence that new methods such as injectables are being
readily accepted by women in the region; these methods
accounted for 62 percent of modern contraceptive users in
Malawi and 66 percent in Ethiopia (National Statistical
Office Malawi and ORC Macro 2005; Central Statistical
Agency and ORC Macro 2006).
Changes in unmet need and satisfied demand
Unmet need measures the gap between demand for family
planning and use of contraception. Expressed as the per-
centage of sexually active women who do not want addi-
tional children but are not using any family planning
method, this measure is often considered a precursor of fer-
tility decline, because it indicates that demand for family
planning services exists but is not being met.
Changes in unmet need can be influenced by a variety of
factors related to fertility preferences or family planning
acceptance, which may or may not be related to the effec-
tiveness of family planning programs. Nevertheless, when
examined in relation to contraceptive prevalence, these
changes provide an estimate of the gap between demand
and utilization of family planning.
Trends in modern CPR and unmet need indicate that in
countries such as Kenya, Madagascar, Malawi, and Zambia,
decline in unmet need has corresponded with an increase in
family planning (figure 25.5) suggesting a convergence of
demand and supply of family planning. In contrast, in other
countries, such as Eritrea, Ghana, Mali, and Senegal, the gap
has remained wide and consistent.
Satisfied demand for contraception is defined as the per-
centage of sexually active women who do not want addi-
tional children and are practicing family planning. Increases
in satisfied demand corresponded with a decrease in the
unmet need for family planning in certain countries in the
region. The percentage change in satisfied demand
increased most in countries such as Madagascar, Mozam-
bique, Tanzania, and Zambia where contraception preva-
lence rates are increasing (figure 25.6).
CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
TRENDS IN TOTAL FERTILITY RATES
Another set of analyses examined trends in total fertility
rates, or the average number of children per woman. In
most countries the TFR declined over time; in some coun-
tries it remained stagnant (figure 25.7). The steepest
declines in average fertility were observed in Ghana, Kenya,
Liberia, Namibia, and Zimbabwe. Other countries, such as
Madagascar, Senegal, and Togo, also showed promising
declines. The TFR increased or remained constant in a few
countries, including Mozambique, Niger, and Nigeria.
Despite increases in contraceptive prevalence in many
countries, fertility decline has been slow. As there tends to
be a time lag between changes in contraceptive use behavior
and a corresponding decline in average fertility, it is likely
that subsequent rounds of surveys will show greater fertility
declines in countries in which CPRs have risen. Fertility
decline also tends to be correlated with demographic and
socioeconomic factors, such as the level of urbanization,
women’s education, women’s labor force participation, and
economic growth. Studies in Africa have shown that differ-
entials in fertility trends across countries are associated with
women’s education, child survival (Kirk and Pillet 1998),
and exposure to modern roles and behaviors linked with
growing urbanization (Garenne and Joseph 2002).
Trends in actual and wanted TFR were examined in
each of the countries for which data were available. Actual
TFR exceeded wanted TFR in all countries (figure 25.8),
indicating that women were not able to regulate their fer-
tility preferences. In countries such as Uganda, that have
high unmet need, the gap between the actual and wanted
TFR widened, suggesting a growing demand for small
family size and the failure of family planning programs to
meet the latent demand for services. Countries, such as
Kenya and Zimbabwe, that have had strong family plan-
ning programs and have improved contraceptive preva-
lence showed a narrowing of the gap between desired and
actual fertility.
Some countries that made progress in the 1990s started to
falter at the beginning of the 2000s. Stagnation on contra-
ceptive prevalence and total fertility rate, was evident (see
figures 25.2, 25.7). Uganda, Tanzania, Malawi, Kenya, and
Ghana underwent large increases in contraceptive preva-
lence in the 1990s but stagnated after 2000. Stagnation in sat-
isfied demand was evident in Kenya, Malawi, Tanzania, and
Zambia. Total fertility rates also stagnated in the 2000s, espe-
cially in Cameroon, Ghana, Kenya, and Zimbabwe. In some
countries, such as Rwanda, Tanzania and Uganda, the gap
between actual and wanted fertility widened after 2000, a
451
452

Figure 25.5 Unmet Contraceptive Need and Modern Contraceptive Prevalence Rate in Selected Countries in Sub-Saharan Africa

Benin Burkina Faso Cameroon Chad Côte d’Ivoire

50

25

Eritrea Ethiopia Ghana Guinea Kenya

50

25

Madagascar Malawi Mali Mozambique Namibia

50
Percent

25

Niger Nigeria Senegal Tanzania Uganda

50

25

0
90

95

00

05

10

90

95

00

05

10

90

95

00

05

10
19

19

20

20

20

19

19

20

20

20

19

19

20

20

20
Zambia Zimbabwe
50

25

0
90

95

00

05

10

90

95

00

05

10
19

19

20

20

20

19

19

20

20

20

Modern contraceptive prevalence rate Unmet need

Source: Demographic and Health Surveys, various years.

 Figure 25.6 Satisfied Demand for Contraception in Selected Countries in Sub-Saharan Africa

Benin Burkina Faso Cameroon Chad Côte d’Ivoire

80
60
40
20

Eritrea Ethiopia Ghana Guinea Kenya

80
60
40
20

Madagascar Malawi Mali Mozambique Namibia

80
Percent

60
40
20

Niger Nigeria Rwanda Senegal Tanzania

80
60
40
20

90

95

00

05

10

90

95

00

05

10
19

19

20

20

20

19

19

20

20

20
Uganda Zambia Zimbabwe
80
60
40
20
90

95

00

05

10

90

95

00

05

10

90

95

00

05

10
19

19

20

20

20

19

19

20

20

20

19

19

20

20

20
East Africa Central Africa Southern Africa West Africa

Source: Demographic and Health Surveys, various years.

453
454

Figure 25.7 Total Fertility Rates in Selected Countries in Sub-Saharan Africa

Benin Burkina Faso Cameroon Chad Côte d’Ivoire

7
6
5
4

Eritrea Ethiopia Ghana Guinea Kenya

7
6
5
4

Liberia Madagascar Malawi Mali Mozambique

Total fertility rate

7
6
5
4

Namibia Niger Nigeria Rwanda Senegal

7
6
5
4

90

95

00

05

10
19

19

20

20

20
Togo Uganda Zambia Zimbabwe
7
6
5
4
90

95

00

05

10

90

95

00

05

10

90

95

00

05

10

90

95

00

05

10
19

19

20

20

20

19

19

20

20

20

19

19

20

20

20

19

19

20

20

20
East Africa Central Africa Southern Africa West Africa

Source: Demographic and Health Surveys, various years.

 Figure 25.8 Actual and Wanted Total Fertility Rates in Selected Countries in Sub-Saharan Africa

Benin Burkina Faso Cameroon Chad Côte d’Ivoire

8
6
4
2

Eritrea Ethiopia Ghana Guinea Kenya

8
6
4
2

Liberia Madagascar Malawi Mali Mozambique

Total fertility rate

8
6
4
2

Namibia Niger Nigeria Rwanda Senegal

8
6
4
2

Tanzania Togo Uganda Zambia Zimbabwe

8
6
4
2
90

95

00

05

10

90

95

00

05

10

90

95

00

05

10

90

95

00

05

10

90

95

00

05

10
19

19

20

20

20

19

19

20

20

20

19

19

20

20

20

19

19

20

20

20

19

19

20

20

20
Actual total fertility rate Wanted total fertility rate

Source: Demographic and Health Surveys, various years.

455
 probable consequence of the weakening of the family plan-
ning program or a shift in desired family size (figure 23.8).
Although conclusive evidence on the reasons for the
stagnation remains elusive, some explanations include
changes in the international policy arena and contextual
changes at the country level. A key policy factor is purported
to be the reduced priority of reproductive health and family
planning after its exclusion from the Millennium Develop-
ment Goals as well as competition for resources from dis-
eases such as tuberculosis, malaria, and HIV (Gillespie
2004). Country-level studies (for example, in Kenya) attrib-
ute the faltering of the fertility decline to dwindling donor
support for family planning and a greater emphasis on
HIV/AIDS and other sexually transmitted diseases (Blacker
and others 2005). A regional study on the stagnation of fer-
tility decline in Eastern Africa concludes that changes in
socioeconomic variables, the family planning program envi-
ronment, and reproductive behavior models are associated
with the decline in contraceptive use and increases in unmet
need, preferences for larger families, and adolescent fertility
(Ezeh, Mberu, and Emina 2009).
The HIV/AIDS epidemic has also impacted fertility lev-
els in Sub-Saharan Africa. The region has the highest preva-
lence of HIV/AIDS and the largest number of people living
with HIV/AIDS in the world. Stagnation in fertility decline
over the past 10 years has been related to the increase in HIV
prevalence. In Zimbabwe, for example, estimated total fer-
tility was 8.5 percent lower than it would have been in the
absence of HIV, and HIV-associated changes in fertility
behavior accounted for one-quarter of the drop in fertility
since the 1980s (Terceira, Simon, and Gregson 2003). In
South Africa, where the prevalence of HIV is among the
highest in the region, the spread of HIV is expected to accel-
erate fertility decline (Moultrie and Timaeus 2003).
DETERMINANTS OF FERTILITY DECLINE
Fertility patterns tend to be influenced by proximate and
socioeconomic determinants of fertility. Findings indicate
that changes on both fronts are taking place in Africa.
The proximate determinants of fertility are the biologi-
cal and behavioral factors through which socioeconomic
and environmental variables operate to influence the rate
of childbearing in a population (Bongaarts 1987). These
determinants have been classified into two broad categories:
fertility-enhancing trends (shortening of breast-feeding
periods and postpartum abstinence, decline in pathological
sterility) and fertility-reducing trends (rise in age at first
union, higher prevalence and effectiveness of contracep-
456 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
tion) (Bongaarts, Frank, and Lesthaeghe 1984). In countries
in which fertility reduction is most pronounced, there is evi-
dence that fertility-reducing variables such as age of mar-
riage have risen (Cohen 1998). However, comparison of the
proximate determinants of fertility in countries in which
the fertility transition is more advanced and those in which
it is delayed indicates that contraceptive use is by far the
most important factor accounting for intercountry differ-
ences (Kirk and Pillet 1998).
Increases in induced abortion—suspected to be a major
method of contraception in urban areas of Africa—are
associated with recent declines in fertility (Garenne and
Joseph 2002). Abortion is the most likely explanation for the
drop in fertility, from 6.9 in 1980 to 5.5 in 2010, in Western
Africa, where contraceptive use remains very low (Cleland,
Ndugwa, and Zulu 2011). An estimated 14 million unin-
tended pregnancies occur in Sub-Saharan Africa every year
(Hubacher, Mavranezouli, and McGinn 2008). Conse-
quently, the demand for medical abortion is expected to be
very high. Because abortion remains illegal in all but a few
countries in the region, women have to seek unsafe abor-
tions from illegal practitioners. It is estimated that more
than 4 million unsafe abortions are performed in Africa
every year (Brookman-Amissah and Moyo 2004). Abortion
is a major risk factor underlying the high levels of maternal
mortality in Africa.
The main socioeconomic determinants of fertility
include socioeconomic status, women’s education, and
urban residence. The negative association between women’s
education and fertility level observed in other settings is
apparent in Africa as well (figure 25.9). Analysis of the rela-
tionship between economic growth and fertility indicates
that increases in GDP are associated with higher rates of
contraceptive prevalence (figure 25.10). The direction and
pathways of causality between fertility and economic
growth remain debatable. In general, socioeconomic change
is believed to modify the incentives to have children, diffuse
new ideas about childbearing through society, and provide
women with better access to contraceptive methods (Bryant
2007). Emerging economic growth prospects in the region
indicate potential for future fertility decline.
Other covariates of fertility, such as infant mortality
rates, have declined in the region. A plot of the relation-
ship between contraceptive prevalence rates and infant
mortality rates indicates the negative correlation between
the two variables (figure 25.11). Fertility decline in coun-
tries such as Botswana, Kenya, and Zimbabwe, which had
lower levels of infant mortality than other countries in the
region, provides evidence that improved rates of child
Figure 25.9 Relationship between Women’s Secondary Education and Contraceptive Prevalence Rates in
Sub-Saharan Africa

50
Y = 7.97 + 0.47X
2 Namibia
R = 0.61
Congo, Rep. of

40 Zimbabwe
Contraceptive prevalence rate (percent)

Swaziland

Madagascar Kenya
30 Zambia
Lesotho
Mozambique
Malawi Cameroon
Tanzania
20 Uganda Congo, Dem. Rep. of Ghana
Benin
Burkina Faso Nigeria
Liberia
Guinea
10 Ethiopia Sierra Leone
Niger
Mali Senegal
Ethiopia
Chad
0
0 20 40 60 80
percent of women with secondary education
Source: Demographic and Health Surveys, 2000–08.

Figure 25.10 Relationship between Annual Percentage Change in GDP and Contraceptive Prevalence Rate in
Sub-Saharan Africa, 1990–2009

100
Contraceptive prevalence rate (percent)

80

60

40

20

0
–5 0 5 10 15
Annual percentage change in GDP

Source: World Bank 2009.

CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED 457
 Figure 25.11 Relationship between Contraceptive Prevalence Rates and Infant Mortality Rates in Sub-Saharan Africa

50 South Africa 1998

Namibia 2006
Congo, Rep. of 2005

40 Zimbabwe 2005
Swaziland 2006
Gabon 2000
Contraceptive prevalence rate (percent)

Zimbabwe 1994

Kenya 2008
30 Madagascar 2008 Zambia 2007
Lesotho 2004
Cameroon 2004
Kenya 1993 Malawi 2004
Mozambique 2003
Rwanda 2007 Togo 1998
Namibia 1992 Tanzania 2004 Burkina Faso 1993
Cameroon 1991 Côte d’Ivoire 1998
20 Uganda 2006 Congo Democratic Republic 2007
Ghana 2008
Ghana 1993
Benin 2006 Côte d’Ivoire 1994
Nigeria 2008 Benin 1996
Comoros 1996Faso 2003 CAR 1994
Burkina
Liberia 2007 Rwanda 1992
Uganda 1995 Madagascar 1992
Ethiopia 2005 Guinea 2005 Zambia 1992
Malawi 1992
10 Sierra Leone 2008
Senegal 2005 Niger 2006 Tanzania 1991
Nigeria 1990 Guinea 1999 Mali 1995
Senegal 1992
Eritrea 2002 Mali 2006 Mozambique 1997
Eritrea 1995 Ethiopia 2000
Mauritania 2000 Niger 1992
Chad 1996
Chad 2004
0

40 60 80 100 120 140

mortality rate

Source: Demographic and Health Surveys, 1986–2009.

survival may be a necessary condition for fertility decline major strides in improving family planning uptake can be
in Africa (Caldwell, Orubuloye, and Caldwell 1992). made if political commitment exists (box 25.3).
Some observers have argued that high rates of fertility
in the region can be linked with the lack of policy level
FAMILY PLANNING POLICIES AND PROGRAMS
commitment for family planning programs. During the
Some studies have tried to identify reasons why certain 1960s and 1970s, African governments were reluctant to
countries in the region underwent fertility decline institute effective family planning programs; political sup-
whereas others did not. One study compares Kenya, where port for family planning in the public sector was weak
total fertility fell about 40 percent between 1980 and 2000, throughout the continent. Since the 1974 and 1984 World
with neighboring Uganda, where fertility declined by Population Conferences, however, governments in several
10 percent. It finds that both economic development and African countries have acknowledged high levels of fertil-
a strong national family planning program were associ- ity and initiated family planning programs (Kalipeni
ated with lower fertility in Kenya (Blacker and others 1995). Africa has lagged other regions on fertility decline
2005) (box 25.1). because family planning programs were introduced rela-
A comparative analysis of Zimbabwe, where the fertility tively late in the region. Family planning programs in
rate fell more rapidly than in Zambia, reveals that a strong Africa are not as strong or as old as those in other parts of
family planning program in Zimbabwe backed by high- the world, but as the experience of many African countries
level political commitment and institutional and financial reveals, if strong and high-quality family planning
stability were key ingredients of success (Lee et al. 1998) programs are developed, people will use them and fertility
(box 25.2). Emerging evidence from Rwanda suggests that will decline (Mbacke 1994). Relatively better progress on

458 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
 Box 25.1 Kenya Case Study
Kenya was the first country in Sub-Saharan Africa to
adopt a population policy, with the launch of a
national family planning program in 1967. Initially,
the promotion of the program was halfhearted and
ineffective: fertility appeared to be rising rather than
falling. In the 1980s, however, the government rapidly
expanded the number of outlets from which contra-
ceptives could be obtained and the number of health
workers trained in family planning. By the end of the
century, more than half of the country’s 3,500 (gov-
ernmental and nongovernmental) health facilities
were offering maternal and child health and family
planning services.
In 1996, the National Council for Population and
Development published its National Population Advo-
cacy and IEC Strategy for Sustainable Development,
1996–2010, which aimed to promote the use of mod-
ern contraceptive methods among less educated
women (Blacker and others 2005). Modern methods
of contraception have been available in Kenya since
1957 through the facilities of the Ministry of Health
and the private sector, including nongovernmental
organizations. Community-based distribution and
Source: Authors.
family planning indicators in Eastern Africa compared to
Western Africa has been attributed to stronger family
planning efforts that ensured wider availability of modern
contraceptive methods (Cleland, Ndugwa, and Zulu 2011).
Although family planning programs in the region have
been weak overall, some encouraging progress in program
implementation began to emerge in the 1980s. A study of
family planning program effort finds that the greatest
improvement among all regions of the world between
1982 and 1989 occurred in Sub-Saharan Africa, where
there was a sharp increase in family planning program
effort indicators, albeit from a low base (Mauldin and Ross
1991). Policy-level support in countries in which strong
commitments existed translated into successful national
family planning programs.
Certain family planning program management strate-
gies have been found to be particularly effective in the
region. Countries such as Kenya and Zimbabwe had strong
family planning associations that spearheaded policy
changes and program implementation. A unified institu-
tional structure responsible for program implementation
was found to be effective in Zimbabwe. In Zambia the
CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
social marketing of contraceptives have been effective
in increasing coverage.
The combined program efforts of public and private
agencies facilitated Kenya’s transformation from the
country with the highest fertility level in the world in
the late 1970s to one in which significant fertility
decline has been achieved. The rate of contraceptive
use among married women increased from 17 percent
in 1984 to 39 percent in 1998, one of the highest rates in
Sub-Saharan Africa (Magadi and Curtis 2003). The main
driving force behind the success was the government’s
effort in increasing the number of family planning service
delivery points and an intensified and focused informa-
tion, education and communication strategy (Aloo-
Obunga 2003). The leveling off in the fertility decline after
2000 may have been caused by problems in the supply of
contraceptives, weaknesses in the quality of care, and
changes in the contraceptive method mix (Pathfinder
International 2005). Another factor may have been the
HIV/AIDS program, which gradually pushed family plan-
ning off the agenda as it became a priority for funding and
strategic programming considerations (Aloo-Obunga
2003; Pathfinder International 2005).
separation of institutions responsible for policy formula-
tion and implementation resulted in a weaker family plan-
ning program (Lee and others 1998).
A few countries have tried community-based distribu-
tion of contraceptives to extend family planning to hard-
to-reach populations, particularly in rural areas. Commu-
nity depots, mobile clinics, women’s groups, and both paid
and volunteer village health workers are some modes of
service delivery used by such programs. Countries such as
Ghana, Kenya, Nigeria, and Zimbabwe have implemented
large community-based distribution programs at the
national level. Although conclusive evidence on the effec-
tiveness of such programs is not available, these programs
provide good examples of successful bottom-up approaches
that have been applied in the region (Phillips, Greene, and
Jackson 1999).
LESSONS LEARNED
Notwithstanding the high levels of aggregate fertility in
Sub-Saharan Africa, some countries in the region have
made significant progress on fertility decline. Ongoing
459
 Box 25.2 Zimbabwe Case Study
After independence, in 1980, the government of Zim-
babwe reformed and expanded the family planning
program with great success. The family planning pro-
gram was spearheaded by the Zimbabwe National Fam-
ily Planning Council, a body created in the early 1980s
and backed by high-level leadership from Sally Mugabe,
the president’s wife, and Ester Boohene, his sister-in-
law. The Zimbabwe National Family Planning Council
built consensus for family planning among opinion
leaders including religious groups, the business com-
munity, mass media, nongovernmental organizations,
and civil servants (Lee, and others 1998).
Initially, Zimbabwe’s family planning program was
clinic-based. A community-based distribution system
was launched in 1983 and it was considered among the
most successful programs of its type in the region
(Way, Cross, and Kumar 1987). Distributors were
selected by the communities they served and paid
government salaries and benefits. They were respon-
sible for making household visits to deliver modern
contraceptives, recruit new acceptors, follow up on
dropouts, and make referrals where necessary (Phillips,
Greene, and Jackson 1999). Mobile clinics covered
about 29 percent of the rural population (Koblinksy
2003). Groups of men and women were recruited to
motivate and educate people in communities about
family planning. Other innovative approaches included
the mobilization of farmers’ wives to provide contra-
ceptives to workers on their farms and the launching
Source: Authors.
transformations in contraceptive use and fertility behav-
ior signal the onset of fertility declines in more countries
in the years to come. Despite tumultuous political situa-
tions several success stories in family planning policy
formulation and program implementation have emerged.
Lessons drawn from countries that have made progress
attest to the importance of political commitment, institu-
tional arrangements, and service delivery strategies in
increasing the use of family planning methods and lower-
ing fertility.
Many countries that were successful in reducing fertility
adopted population policies and instituted family planning
programs relatively early. Programs in Botswana, South
Africa, and Zimbabwe have been considered particularly
successful in this regard (Lucas 1992). High-level policy
commitment and political ownership of the population
program was a key ingredient for success. Political com-
460 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
of a large national information, education, communi-
cation campaign that promoted family planning with
messages targeting men. The number of service deliv-
ery points was increased, particularly in rural areas;
the number of family planning personnel more than
doubled in some units; and the government made
health care free to lower-income groups, thereby
removing a major barrier to contraceptive use (Zinaga
1992).
The availability and quality of family planning and
health services in the community was a key determi-
nant of higher rates of adoption of modern contra-
ceptives (Thomas and Muvandi 1994). The impact of
community-based distributors was associated with
increased adoption of modern methods. Mobile fam-
ily planning clinics had a powerful impact on adop-
tion, as did the presence of a general hospital in the
area. These two investments in infrastructure had an
above-average impact on women with little education.
The program also provided a range of contraceptive
methods, including Norplant, the female condom, and
emergency contraception (Koblinksy 2003). Family
planning was positioned as an integral part of the
maternal and child health program. The primary
health care strategy adopted by the government
included both maternal and child health and family
planning. All service delivery units were instructed to
provide family planning as an integral part of their
maternal and child health services (Zinaga 1992).
mitment is not enough, however; the leadership must pro-
vide contraceptives and appropriate outlets for obtaining
them and create an environment that is conducive to
adoption of family planning (Caldwell, Orubuloye, and
Caldewell 1992).
The existence of strong family planning programs is a
prerequisite to reducing fertility. Family planning pro-
grams that have delivery points throughout the country;
provide a range of contraceptive methods; ensure easy
availability of contraceptives; adopt a reproductive health
approach; and reach adolescents, men, and unmarried peo-
ple are most likely to accelerate progress toward fertility
decline in Africa (Caldwell and Caldwell 2002). Some spe-
cific service delivery strategies that have been found to be
effective in Africa are those that promote spacing methods,
give women the means to assume responsibility over con-
traceptive adoption, and allow women to use contraception
 Box 25.3 Rwanda Case Study
In the 1980s Rwanda had one of the highest levels of
fertility in the world (8.6). The availability of family
planning service was very limited, and cultural atti-
tudes dictated the desire for a large family. As a result,
contraceptive prevalence rate was extremely low, at
3–4 percent in 1988 (May, Mukamanzi, and Vekemans
1990).
The National Office of Population (ONAPO) was
established in 1981 to increase demand for and improve
the supply of family planning services. It made some
initial progress, but its gains were lost during the geno-
cide of 1994. After the recovery, the government intro-
duced sweeping reforms that have resulted in dramatic
progress on health and family planning indicators,
with the contraceptive prevalence rate increasing
several-fold, skyrocketing from 4.3 percent in 2000 to
26.1 in 2007.
The importance of providing family planning ser-
vices as a rationale for national development found
strong support at the top leadership level of the govern-
ment. A number of innovative strategies were applied to
Source: Authors.
without their partners’ knowledge (Caldwell, Orubuloye,
and Caldwell 1992).
Policies that go beyond simply increasing contraceptive
prevalence to address the proximate determinants of fertil-
ity can accelerate fertility decline in significant ways.
Increasing the age of marriage through legislation and
behavior change, encouraging natural child spacing
through promotion of exclusive breastfeeding, and reducing
the risk of unsafe abortion by removing legal restrictions
will be key factors underlying fertility decline in Africa
(Guengant and May 2002). Improvements in health, educa-
tion, and socioeconomic factors can also catalyze fertility
reduction in the region. For example, experiences from
Botswana, Kenya, and Zimbabwe show that countries that
are making improvements in child survival and women’s
education have the greatest potential for reducing fertility
(Caldwell, Orubuloye, and Caldwell 1992).
Although fertility transition has begun in Africa, it is lim-
ited to a few countries; fertility decline across the region
remains a distant goal. Even in countries in which contra-
ceptive prevalence is increasing, huge differentials exist by
socioeconomic strata, urban-rural residence, and correlates
such as female education and autonomy. The stagnation in
recent years provides compelling evidence that faltering
CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
improve access and quality of health services. Some
successful strategies included the introduction of a
performance-based financing mechanism at various
levels. Performance contracts were established
between the presidency and district mayors (Imihigo)
that included an indicator on family planning. Perfor-
mance incentives were also given to health workers to
improve their motivation levels and to health centers
for providing quality care. Universal health insurance
schemes (mutuelles de santé) further enhanced the
coverage of health care and encouraged community
involvement in health provision. Other strategies
included decentralizing health services, strengthening
contraceptive supply systems, training health workers
for family planning provision, and establishing sec-
ondary posts as an alternate means for providing
modern contraceptives to circumvent that religiously
affiliated health facilities were not allowed to do so
(Solo 2008). Rwanda’s remarkable turnaround from
its conflict-torn past to current achievements serves
as a model for other countries in the region.
policy and resource commitments can easily reverse the
gains that have been made in the past.
Family planning remains an unfinished agenda in the
region, because high fertility and rapid population growth
present a great threat to the achievement of the Millennium
Development Goals (Cleland and others 2006). A greater
thrust in this direction will be required to sustain and
improve the prospects for health and development in Sub-
Saharan Africa in the coming decades.
REFERENCES
Aloo-Obunga, C. 2003. Country Analysis of Family Plan-
ning and HIV/AIDS: Kenya. Policy Project, Washington,
DC.
Blacker, J., C. Opiyo, M. Jasseh, A. Sloggett, and J. Ssekamatte-
Ssebuliba. 2005. “Fertility in Kenya and Uganda: A Com-
parative Study of Trends and Determinants.” Population
Studies 59 (3): 355–73.
Bongaarts, J. 1987. “The Proximate Determinants of Fertil-
ity.” Technology in Society 9 (3–4): 243–60.
Bongaarts, J., O. Frank, and R. Lesthaeghe. 1984. “The Prox-
imate Determinants of Fertility in Sub-Saharan Africa.”
Population and Development Review 10 (3): 511–37.
461
 Brookman-Amissah, E., and J. B. Moyo. 2004. “Abortion
Law Reform in Sub-Saharan Africa: No Turning Back.”
Reproductive Health Matters 12 (24): 227–34.
Bryant, J. 2007. “Theories of Fertility Decline and the Evi-
dence from Development Indicators.” Population and
Development Review 33 (1): 101–27.
Caldwell, J. C., and P. Caldwell. 1987. “The Cultural Context
of High Fertility in Sub-Saharan Africa.” Population and
Development Review 13 (3): 409–37.
———. 1988. “Is the Asian Family Planning Program
Model Suited to Africa?” Studies in Family Planning
19 (1): 19–28.
———. 2002. “Africa: The New Family Planning Frontier.”
Studies in Family Planning 33 (1): 76–86.
Caldwell, J. C., I. O. Orubuloye, and P. Caldwell. 1992.
“Fertility Decline in Africa: A New Type of Transition?”
Population and Development Review 18 (2): 211–42.
Central Statistical Agency, and ORC Macro. 2006. Ethiopia
Demographic and Health Survey 2005. ORC Macro,
Calverton, MD.
Cleland, J., S. Bernstein, A. Ezeh, A. Faundes, A. Glasier, and
J. Innis. 2006. “Family Planning: The Unfinished
Agenda.” The Lancet 368 (9549): 1810–27.
Cleland J., R. P. Ndugwa, and E. M. Zulu. 2011. “Family
planning in sub-Saharan Africa: Progress or Stagna-
tion.” Bulletin of the World Health Organization 89:
137–143.
Cohen, B. 1998. “The Emerging Fertility Transition in Sub-
Saharan Africa.” World Development 26 (8): 1431–61.
Ezeh, A. C., B. U. Mberu, and J. Emina. 2009. “Stall in Fer-
tility Decline in Eastern African Countries: Regional
Analysis of Patterns, Determinants and Implications.”
Philosophical Transactions of the Royal Society Biological
Sciencies 364 (1532): 2991–3007.
Garenne, M., and V. Joseph. 2002. “The Timing of the Fer-
tility Transition in Sub-Saharan Africa.” World Develop-
ment 30 (10): 1835–43.
Gillespie, D. G. 2004. “Whatever Happened to Family
Planning and for That Matter, Reproductive Health?”
International Family Planning Perspectives 30 (1):
34–38.
Guengant, J. P., and J. M. May. 2002. Impact of the Proximate
Determinants on the Future Course of Fertility in Sub-
Saharan Africa: Prospects of Fertility Decline in High Fer-
tility Countries. New York: United Nations.
Hubacher, D., I. Mavranezouli, and E. McGinn. 2008.
“Unintended Pregnancy in Sub-Saharan Africa: Mag-
nitude of the Problem and Potential Role of Contra-
ceptive Implants to Alleviate It.” Contraception 78 (1):
73–78.
462 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED
Kalipeni, E. 1995. “The Fertility Transition in Africa.” Geo-
graphical Review 85 (3): 286–300.
Kirk, D., and B. Pillet. 1998. “Fertility Levels, Trends, and
Differentials in Sub-Saharan Africa in the 1980s and
1990s.” Studies in Family Planning 29 (1): 1–22.
Koblinksy, M. 2003. Reducing Maternal Mortality: Learning
from Bolivia, China, Egypt, Honduras, Indonesia, Jamaica,
and Zimbabwe. Washington, DC: World Bank.
Lee, K., L. Lush, G. Walt, and J. Cleland. 1998. “Family Plan-
ning Policies and Programmes in Eight Low-Income
Countries: A Comparative Policy Analysis.” Social Science
& Medicine 47 (7): 949–59.
Lucas, D. 1992. “Fertility and Family Planning in Southern
and Central Africa.” Studies in Family Planning 23 (3):
145–58.
Magadi, M. A., and S. L. Curtis. 2003. “Trends and Determi-
nants of Contraceptive Method Choice in Kenya.” Stud-
ies in Family Planning 34 (3): 149–59.
Mauldin, W. P., and J. A. Ross. 1991. “Family Planning Pro-
grams: Efforts and Results, 1982–89.” Studies in Family
Planning 22 (6): 350–67.
May, J. F., M. Mukamanzi, and M. Vekemans. 1990. “Family
Planning in Rwanda: Status and Prospects.” Studies in
Family Planning 21 (1): 20–32.
Mbacke, C. 1994. “Review: Family Planning Programs and
Fertility Transition in Sub-Saharan Africa.” Population
and Development Review 20 (1): 188–93.
Moultrie, T. A., and I. M. Timaeus. 2003. “The South African
Fertility Decline: Evidence from Two Censuses and a
Demographic and Health Survey.” Population Studies
57 (3): 265–83.
National Statistical Office Malawi, and ORC Macro. 2005.
Malawi Demographic and Health Survey 2004. ORC
Macro, Calverton, MD.
Pathfinder International. 2005. Community-Based Family
Planning in Kenya: Meeting New Challenges. Nairobi and
Watertown, MA.
Phillips, J. F., W. L. Greene, and E. Jackson. 1999. Lessons
Learned from Community-Based Distribution of Family
Planning in Africa. No. 121, Population Council,
New York.
Solo, J. 2008. “Family Planning in Rwanda: How a Taboo
Topic Became Priority Number One.” IntraHealth Inter-
national, Chapel Hill, NC.
Terceira, N., G. Simon, and S. Gregson. 2003. “The
Contribution of HIV to Fertility Decline in Rural
Zimbabwe, 1985–2000.” Population Studies 57 (2):
149–64.
Thomas, D., and I. Muvandi. 1994. “The Demographic
Transition in Southern Africa: Another Look at the
 Evidence from Botswana and Zimbabwe.” Demography
31 (2): 185–207.
UNDESA (United Nations Department of Economic and
Social Affairs). 2008. World Population Prospects, the
2008 Revision. New York: United Nations.
United Nations Population Division. 2009. World Contra-
ceptive Use. Nw YorK; United Nations Department of
Economic and Social Affairs.
Way, A. A., A. R. Cross, and S. Kumar. 1987. “Family Plan-
ning in Botswana, Kenya and Zimbabwe.” International
Family Planning Perspectives 13 (1): 7–11.
World Bank. 2009. World Development Indicators. Washing-
ton, DC: World Bank.
Zinaga, A. F. 1992. “Development of the Zimbabwe Family
Planning Program.” Policy Research Working Paper,
World Bank, Washington, DC.

CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED 463
 CHAPTER 26
Achieving Universal Primary Education
through School Fee Abolition:
Some Policy Lessons from Uganda
B. Essama-Nssah
n 1997 the government of Uganda abolished school
I
fees for primary education. Immediately, primary
enrollment rates skyrocketed, with gross enrollment
rates rising from 77 percent in 1996 to 137 in 1997 and net
enrollment rising from 57 percent to 85 percent.1 Enroll-
ment remained high over the following decade, with poor
children, girls, and rural residents benefiting dispropor-
tionately from the increase in access.
The quantitative success of the government’s universal
primary education (UPE) initiative put significant stress on
the country’s educational infrastructure, however, with a
consequent toll on the quality of primary schooling.
Increased enrollment led to overcrowding, multiple shifts,
shortages of teachers and material, and an increase in over-
age students.2 This chapter analyzes the outcomes of UPE in
Uganda in order to draw policy lessons from the experience.
THE NEED FOR REFORM
Uganda’s education sector achieved considerable progress fol-
lowing independence, but that progress was reversed as a
result of various sociopolitical crises in the 1970s and 1980s.
In 1985 the level of government expenditure on education was
about 27 percent that of the 1970s (Appleton 2001). Partly as
a result of the high costs of education to families, the gross pri-
mary enrollment rate stood at 50 percent in 1980—the same
rate as in 1960. Significant improvement was observed in
1985, when the gross primary enrollment rate increased to
73 percent, but it remained stagnant at that level until 1995.
In 1996 the government committed itself to providing
UPE. It abolished school fees at the primary level starting in
January 1997. The policy was not adopted overnight. It is an
integral part of a broader reform program, which can be
traced back to the creation in 1987 of the Education Policy
Review Commission (see discussion later in chapter).
Education contributes directly to human development
and is an investment in economic prosperity. In recogni-
tion of the fact that it has both intrinsic and instrumental
value for development, the government placed education at
the center of its Poverty Eradication Action Plan (PEAP).
The current 2010–15 five-year National Development Plan
also recognizes the importance of education for sustained
economic growth and social transformation.
OUTCOMES OF REFORM
The goal of Uganda’s primary education policy is to ensure
that every child enrolls at the appropriate age and success-
fully completes the full cycle of education.3 Achievement of
this objective requires expanding educational opportunities
and improving educational outcomes. The review of the
results of reform presented in this section focuses on four
outcomes: school enrollment and its cost, age at enrollment,
completion rates, and equity considerations.
School enrollment
Between 1996 (the year of the policy announcement) and
1997 (the first year of implementation), primary school
465
 enrollment in Uganda increased from about 3.1 million
(a gross enrollment rate of 77 percent) to 5.3 million (a gross
enrollment rate of 137 percent), a 58 percent increase. The
net enrollment rate rose from 57 percent in 1996 to 85 per-
cent in 1997 and 90 percent in 1999 (MOES 1999).
To what extent are these changes in enrollment a result
of the UPE policy? The 1997 education policy reform can
be viewed as a natural experiment in which a sharp change
in government policy is an exogenous variation in a treat-
ment variable to which only a segment of the population
(that is, primary school students) is exposed. Deininger
(2003) estimates a linear probability model for primary
school enrollment among three age groups: 6–12, 6–8, and
9–12 (see annex 26A for technical details). To control for
unobserved effects of factors associated with a common
economic environment, he estimates the same equation
for secondary school enrollment (the 12–18 age group).
All primary school regressions show a positive and signif-
icant time trend, implying that the probability of primary
school enrollment increased for everyone in the sample.
When all other variables are set to their mean levels, the
probability of a child being enrolled in primary school in
1999 is about 60 percent higher than it was in 1992.
Deininger finds no significant effect of UPE on secondary
school enrollment. Essama-Nssah, Leite, and Simler
(2008) replicate this analysis using data for 1992 and 2005.
They find similar results.
Figure 26.1 Boys’ and Girls’ Enrollment in Public Primary Schools in Uganda, 2000–10
6
Millions
2
2000 2001 2002 2003
Total
Source: MOES 2010.
466 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
The surge in enrollment associated with the reduction in
the cost of primary schooling is evidence that this cost was
a significant impediment to primary school enrollment by
the poor (Deininger 2003). Indeed, lack of interest and the
cost of enrollment are the two major determinants of
nonenrollment.
As a result of these reforms, the importance of cost as a
reason for not enrolling in school declined significantly: in
1992, 42 percent of households cited cost as a factor. This
figure fell to just 13 percent by 2005 (Essama-Nssah, Leite,
and Simler 2008). This constraint did not change in urban
areas, and it declined by less in the central region, where it
fell from about 46 percent in 1992 to about 36 percent in
2005, than elsewhere in Uganda. Lack of interest, which
increased from 49 percent in 1992 to 56 percent in 2005,
remains a severe constraint to enrollment.
To what extent have increases in enrollment been sus-
tained over time? Total enrollment increased steadily, from
5.4 million in 2000 to about 7.4 million in 2010 (figure
26.1). The gross enrollment ratio dropped from about 128
percent in 2000 to about 104 percent in 2004 (figure 26.2).
After a slight increase in 2005, it has been hovering around
115 percent since 2006. After peaking at 100 percent in 2003,
the net enrollment ratio has shown an upward trend, rising
from 90 percent in 2004 to 96 percent in 2010. These trends
suggest that the sustained efforts of the government to
improve education policy led to and maintained an increase
2004 2005 2006 2007 2008 2009 2010
Girls Boys
Figure 26.2 Gross and Net Primary Enrollment Rates in Uganda, 2000–10

140

130

120
Percent

110

100

90

80
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Gross primary enrollment rate Net primary enrollment rate

Source: MOES 2010.

Note: The gross primary enrollment rate (GER) is the ratio between the number of children enrolled in primary school and the number of children of pri-
mary school age. The net primary enrollment rate (NER) is the ratio between the number of children of primary school age enrolled in primary school
and the number of children of primary school age. See note 1 for more information.

in enrollment not only for over-age but for under-age chil-
dren as well.
Essama-Nssah, Leite, and Simler (2008) study the impact
of the policy shift on the cost of enrollment using a linear
regression model with interactive terms (analogous to the
linear probability model discussed above). The dependent
variable is the logarithm of school fees paid (as reported in
the 1992 and 2005 surveys). The specification accounts for
the fact that school fees are observed only for students who
are enrolled and not for those who are not enrolled in
school. This potential selection bias calls for the application
of Heckman’s (1976) two-step estimator. This approach,
which is analogous to instrumental variable estimation,
requires simultaneous modeling of both exposure to policy
intervention and the associated outcome. Identification
therefore hinges on an exclusion restriction—that is, the
variation of at least one exogenous variable that affects par-
ticipation (enrollment) but not the outcome (fees paid).
Essama-Nssah, Leite, and Simler apply this methodology
to both the primary and secondary levels of education, using
community-level costs and peers’ enrollment rates in the
identifying restriction. Thus at the household level, the con-
struction of these variables excludes the information pertain-
ing to the household under consideration. This construction
ensures that the instrumental variables used for identification
are not directly related to the outcome variable. They find
that over time there has been a significant increase in school
fees paid by families at all levels of schooling. However, stu-
dents in public primary schools were paying less in 2005 than
they paid in 1992, suggesting that UPE may have reduced the
cost of enrolling in public primary schools. A comparison of
the policy impact at the primary and the secondary levels sug-
gests that the policy shift may indeed have moderated the
increase in school fees at the primary level but that there was
no spillover effect to the secondary level. Their results show
that in 2005, fees paid at the secondary level of education
were higher in public than in private schools. UPE did lower
the cost of primary education.
Age at enrollment and completion rates
Deininger (2003) finds that mother’s education is the most
important factor influencing timely enrollment (that is,
enrollment between the ages of 6 and 8). It does not affect
children who enroll between the ages of 9 and 12.
Grogan (2009) uses data from the 1995 and 2000 Demo-
graphic and Health Surveys (DHS) along with the 2001
DHS EdData survey for Uganda to assess the impact of UPE
on the likelihood of enrolling in school before the age of
nine—a key factor in preventing drop-out.4 She shows that
starting school before age nine is associated with a 16–26
percent increase in the probability of completing at least

CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION 467
 seven years of schooling. Using the regression discontinuity
framework from the program evaluation literature, she
shows that elimination of school fees associated with the
UPE policy had a positive albeit small effect on the propen-
sity to enroll in primary school before age nine (the overall
effect is estimated at 3 percent). This finding is consistent
with the conclusion reached earlier, on the basis of the evo-
lution of enrollment rates, that UPE led to an increase in the
enrollment of under-age children.
Essama-Nssah, Leite, and Simler (2008) analyze the
impact of the policy shift on primary school completion. In
particular, they focus on the probability of completing the
seventh year of primary school (P7), the last grade of pri-
mary school in Uganda. As in the case of enrollment, house-
hold income, age, parental education, and urban residence
are key determinants of the likelihood of completing pri-
mary school. These factors have a positive and significant
effect on the probability of completing P7. Essama-Nssah,
Leite, and Simler find that the policy shift may not have
improved the chances of completing P7. They also find that
the income constraint on the chances of completing P7 has
become more severe.
The overall completion rate fell continuously between
2004 and 2007, from 62 percent to 47 percent (figure 26.3).
This trend seems to support the results of the analysis based
on the 2005 household survey discussed earlier. The com-
pletion rates for girls and boys follow the same pattern, with
Figure 26.3 Boys’ and Girls’ Primary School Completion Rates in Uganda, 2001–10
75
70
65
60
Percent
55
50
0
2001 2002 2003 2004
Overall
Source: MOES 2010.
468 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
rates for boys consistently above those for girls. Improving
completion rates thus remains a challenge to policy makers.
Equity
The achievement of equity is increasingly seen as an impor-
tant goal of socioeconomic development. Identifying win-
ners and losers is therefore an important concern for policy
makers. Doing so depends fundamentally on the response
of socioeconomic agents to the incentives of the policy
under consideration.
The distributional implications of an intervention
depend on the variation in the impacts across groups or
individuals. Such impact heterogeneity stems from both
treatment heterogeneity (if the dose received depends on
some unit attributes) and heterogeneity in individual cir-
cumstances (which determine the response to treatment).
The same heterogeneity that must be controlled for in iden-
tifying causal effects must be accounted for in identifying
winners and losers. Models that include interaction between
the treatment indicator and covariates offer a framework for
the analysis of systematic variation in mean impacts across
socioeconomic groups. Interactive models focus on differ-
ences in subgroup means.
The regression framework used by Deininger (2003) is
suitable for the analysis of the distributional implications of
the UPE policy. In 1992, before the introduction of UPE,
2005 2006 2007 2008 2009 2010
Boys Girls
there was a strong bias against girls in Uganda, and parental
income had a strong impact on the probability of enrollment.
The probability of enrollment, which increases with parental
income, was 5 percent higher for boys than for girls. The rela-
tionship between parental education and the probability of
primary school enrollment was also positive and significant.
Fixing all variables at their mean level and increasing the
father’s education by one year would lead to an increase in the
probability of enrollment of about 3 percent; an additional
year of education by the mother would increase the probabil-
ity of her child enrolling in school by 4 percent. The baseline
results show significant regional disparities.
Interacting the policy variable with a variety of house-
hold characteristics, Deininger (2003) shows that UPE has
been pro-poor. The positive impact of parental income on
the probability of enrollment is significantly lower after
reform. Comparing 1992 and 2005, Essama-Nssah, Leite,
and Simler (2008) find a significant increase in enrollment
of the poor. At the bottom quintile, enrollment increased by
more than 28 percentage points (from 50.2 percent in 1992
to 78.8 percent in 2005). However, the gap between the bot-
tom and the top quintile does not seem to have narrowed.
The gap between boys and girls that was evident in 1992 dis-
appeared in 2005, with the gap in the gross enrollment rate
for girls (23.5 percent) somewhat higher than that for boys
(20.4 percent). These results confirm the trends observed by
Deininger (2003).
Figure 26.4 Boys’ and Girls’ Success Rates on Uganda’s Primary School Leaving Examination, 2000–10
95
90
85
80
Percent
75
70
65
60
2000 2001 2002 2003
Overall
Source: MOES 2010.
CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
The achievement gap between boys and girls narrowed as
well. In 2001 the primary school completion rate was 71
percent for boys and 63 percent for girls. The two groups
almost achieved parity in 2008, when the completion rate
was 51 percent for boys and 49 percent for girls. For 2010
the gap is estimated at 6 percentage points.
Convergence in achievement by boys and girls is evident
from figure 26.4, which shows the success rate on the Pri-
mary School Leaving Examination over the past decade. In
2000, 90 percent of male candidates but only 63 percent of
female candidates passed the examination. In 2010, 92 per-
cent of boys and 90 percent of girls who take the exam were
expected to pass.
Another dimension of gender inequality deserves special
attention. The 2010–15 five- year national development
plan notes that the largest proportion of out-of-school chil-
dren are girls. Girls are also more likely than boys to drop
out of school or repeat grades.
The World Development Report 2006 argues for the pur-
suit of equity on both intrinsic and instrumental grounds. It
defines equity in terms of a level playing field on which indi-
viduals have equal opportunity to freely pursue chosen life
plans and are spared from extreme deprivation in outcomes.
The equitable distribution of educational resources is one of
the best ways to try to equalize opportunity across socioe-
conomic groups. Observed inequality of outcomes among
groups defined on the basis of circumstances beyond their
2004 2005 2006 2007 2008 2009 2010
Boys Girls
469
 control is one potential indicator that current policies have
not fully achieved equality of opportunity.
Regional differences in resource allocations are wide in
Uganda (figure 26.5). Of the country’s four regions, the
eastern and the northern have the highest pupil-teacher and
pupil-classroom ratios. These data suggest that children in
these regions do not have the same learning opportunities
as their counterparts in the central and western regions.
POLICY FRAMEWORK
The abolition of primary school fees and the associated
measures adopted in the context of the UPE reform led to a
significant and lasting expansion of educational opportuni-
ties and an improvement in equity in Uganda. The policy
process that underpins these achievements has been held up
as a leading example of the crucial role played by country
ownership and donor cooperation within a sectorwide
approach to policy reforms.
The success or failure of a policy is usually assessed in
terms of its objectives. But to understand the observed out-
comes one needs to consider the policy-making process.
The UPE reform sought to expand educational opportuni-
ties and improve teaching and learning outcomes. There is
evidence that the gains in access and equity have not been
fully matched by improvement in educational outcomes.
This section considers the policy framework in order to
identify key determinants of these outcomes and draw pol-
icy lessons from the Ugandan experience.
Figure 26.5 Pupil-Classroom and Pupil-Teacher Ratios in
Uganda, by Region, 2008
100
90
80
Percent
70
60
50
40
30
Central East North West
Pupil-classroom ratio Pupil-teacher ratio
Source: MOES 2008.
470 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
Vision and political commitment
A clear strategic vision focused on the role of education in
poverty reduction as well as strong political commitment
at various levels of government backed by a sound policy
framework underpinned the “big bang” approach adopted
in Uganda and contributed to its success. In 1996–97 the
top political leadership in Uganda showed its commitment
to poverty reduction by spearheading the formulation of
the Poverty Eradication Plan. Uganda was among the first
low-income countries to prepare a comprehensive and
participatory national strategy for poverty reduction; its
experience inspired the design of the Poverty Eradication
Plan in Uganda and the Poverty Reduction Strategy Pro-
gram (PRSP) approach (Mackinnon and Reinikka 2000).
The government placed education at the center of this
action plan, in recognition that education has both intrin-
sic and instrumental value for development. To show its
determination, the government quickly translated the
action plan into a budget and medium-term expenditure
framework.
Nature and scope of education policy reform
The World Development Report 2004 (World Bank 2003)
argues that the effectiveness of social services in developing
human capital depends fundamentally on the method of
delivery and the behavior of key actors, including policy
makers, service providers, and potential beneficiaries.5 Out-
comes are thus jointly determined by supply- and demand-
side factors and the interactions among them. Supply-side
interventions seek to increase the level and quality of ser-
vices provided. Such interventions usually entail building
and staffing facilities, providing inputs, implementing insti-
tutional reforms, and strengthening the incentives facing
services providers. On the demand side, typical interven-
tions seek to improve the ability of participants to benefit
from the service provided. Some interventions include
incentives for potential beneficiaries to provide the requisite
level of effort to achieve the desired outcome. One way of
stimulating demand for services is to make them more
affordable. Other incentive-based interventions make
resource transfers conditional on some desired behavior.
For example, a conditional transfer program may seek to
promote human capital accumulation by making cash or in-
kind transfers to poor families provided that their school-
National
age children stay in school and young children and pregnant
and nursing women participate in some health-enhancing
activities.
 The education policy reform in Uganda is broadly consis-
tent with this framework. The sudden increase in the demand
for public primary education created a series of challenges
for the government, related mainly to financing the reform,
improving the quality of primary education, ensuring
equitable access to primary education, and planning beyond
primary school. To improve the quality of primary educa-
tion, the government has focused on five areas: curriculum
development, provision of basic learning materials, teacher
development, language of instruction for lower primary
pupils, and establishment and maintenance of standards.
It took some time for reform in these areas to signifi-
cantly improve the quality of teaching and learning in
Uganda. Test results from the National Assessment of
Progress in Education between 1996 and 2000 showed
deterioration in student performance in math, reading, sci-
ence, and social studies (Bategeka 2005). The key factors
explaining the decline outcome were lack of coherence and
consistency within the system and changes in teaching and
learning methods (Penny and others 2008).
Over time the situation has improved, for both boys and
girls. Between 2003 and 2010, the overall numeracy rate for
students in P3 rose from 42 percent to 72 percent, with sim-
ilar increases for girls and boys (figure 26.6). These gains
were more dramatic than were gains in literacy (figure 26.7).
Uganda uses two types of grants to increase equitable
access to primary education: capitation grants and school
Figure 26.6 Numeracy Rates for Boys and Girls in P3 in Uganda, 2003–10
75
70
65
60
Percent
55
50
45
40
35
2003 2004 2005
Overall
Source: MOES 2010.
CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
facilities grants. The purpose of the capitation grant is to
shift some of the burden of school fees from parents to the
government and to provide schools with some of the
resources necessary to run the school and support teaching
and learning.6 The capitation grant is meant to cover tuition
only; families remain responsible for writing materials, uni-
forms, and lunches. This grant is paid to schools in nine
monthly installments, at an annual rate of U Sh 5,000
(about $3.00) per pupil in P1–P3 and U Sh 8,100 (about
$4.75) per pupil in P4–P7 (Penny and others 2008).7
The school facilities grant is designed to assist schools in
the neediest communities in building classrooms, latrines,
and teachers’ houses and procuring furniture. After a favor-
able review in 1999, it became the only mechanism through
which public funds are channeled for the construction of
school facilities. Both grants are conditional, to the extent
that funding is given to districts or municipalities under
strict guidelines and regulations and under the supervision
of the Ministry of Education and Sports (MOES).8
Planning
The success of a policy initiative like Uganda’s UPE requires
background analytical work, identification of financing
sources, and the development of implementation and
monitoring capacity. Analytical work could take the form
of an overall assessment of the performance of the current
2006 2007 2008 2009 2010
Boys Girls
471
 Figure 26.7 Literacy Rates for Boys and Girls in P3 in Uganda, 2003–10
60
55
50
Percent
45
40
35
30
2003 2004 2005
Overall
Source: MOES 2010.
system. Such an assessment would identify constraints that
must be addressed and consider the feasibility and desir-
ability of a variety of policy options. In an environment in
which parents value education, as is the case in Uganda,
abolition of school fees will normally lead to a surge in
enrollment. It is therefore important to assess the expected
increase in demand for schooling and plan for the addi-
tional resources (teachers, facilities, and teaching and learn-
ing material) needed to cope with the surge.
The UPE reform in Uganda followed the “big bang”
approach to policy reform, but adoption of the reform fol-
lowed a long gestation period, and it benefited from many
critical prior actions, starting with the creation of the Edu-
cation Policy Review Commission in 1987. In its 1989
report to the government, the commission set an ambitious
goal for primary education, recommending that education
policy ensure that every child enroll in school at the appro-
priate age and successfully complete the full cycle (P1–P7)
(Grogan 2009).9 The first response to this recommendation
came in 1993, in the form of the Primary Education and
Teacher Development Project, supported by the World Bank
and the U.S. Agency for International Development
(USAID). The project, which sought to introduce a coun-
trywide evaluation framework for overall progress in educa-
tion, led to the establishment of a Teacher Development and
Management System and the creation of an Instructional
Material Unit. Under the Teacher Development and Man-
agement System, all public schools (also known as govern-
ment-aided schools) were organized into clusters of 18
472 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
2006 2007 2008 2009 2010
Boys Girls
schools. One school in each cluster serves as the coordi-
nating center school, in the sense that it had a trained
teacher (the outreach tutor) in charge of helping parents,
community leaders, teachers, and principals in each school
in the cluster coordinate their efforts for improving teach-
ing and learning outcomes. Outreach tutors are supervised
and supported by outreach administrators from core pri-
mary teacher colleges. By 1998, 18 core primary teacher
colleges were supervising about 550 outreach tutors
(MOES 1999).
In 1994 the National Curriculum Development Center
lost its monopoly on the production of instructional mate-
rials. The role of the Instructional Material Unit is to help
districts acquire instructional materials (including books)
and to ensure that the process is consistent with government
regulation designed to ensure transparency.
In 1995 a monitoring and evaluation program, the
National Assessment of Progress in Education was set up
to attempt to determine the amount of and type of knowl-
edge acquired by pupils relative to the objectives of the
curriculum.
Partnerships
The outcome of a policy intervention hinges critically on
individual behavior and on the rules that govern interaction
among stakeholders. In the context of the provision of pub-
lic services, the World Development Report 2004 argues that
there is a need for a governance framework that defines
accountability for results between policy makers and
providers, between policy makers and beneficiaries, and
between providers and beneficiaries. The UPE reform was
embedded in Uganda’s national Poverty Eradication Plan,
which is consistent with the key principles underlying the
PRSP approach of accountability based on partnership. In
the context of education reform, now a strategic component
of the National Development Plan, the central government
is engaged in a two-tier partnership involving development
partners and domestic stakeholders.
The government of Uganda and its development part-
ners adopted a sectorwide approach to deal with resource
mobilization and aid coordination. The common vision
binding this partnership stems from both the adoption of
the PRSP approach toward the end of the 1990s and the
principles voiced at the World Education Forum held in
Dakar in 2000. The government agreed with its develop-
ment partners that 31 percent of the recurrent discretionary
budget would go to education and that at least 65 percent of
that amount would be allocated to primary education (Hig-
gins and Rwanyange 2005). The World Education Forum
led to a commitment by the international community to
support any developing country that seriously engaged in
the pursuit of policies aimed at securing quality education
for all (Eilor 2004).
The sectorwide approach started with the Education
Strategic Investment Plan (ESIP) 1998–2003. At the end
of the ESIP funding agreement, the Medium-Term Budget
Framework became the common budget support modal-
ity (Penny and others 2008). The sectorwide approach
emphasized actions designed to enhance local leadership
and the integration of development partner and govern-
ment efforts. Budget support was thus conditioned on a
few key outputs, and outcomes related fiduciary integrity
and progress toward improving service delivery and equi-
table access. The Heavily Indebted Poor Country Initia-
tive led to an increase in funds available to the education
sector. The effectiveness of this institutional arrangement
was significantly enhanced by the use of sector reviews
and budget working groups. These mechanisms, along
with an effective and reliable Education Management
Information System, are now playing a fundamental role
in the education planning and budgeting processes
(Penny and others 2008).
The partnership between the government of Uganda and
the aid community is often cited as a success story. This
partnership is based on a common vision about the effec-
tiveness of poverty reduction strategies. It pursues a holistic
approach, putting the country in the driver’s seat. Both sup-
CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
ply and demand constraints to service delivery are
addressed within a common sector policy and planning
framework supported by well-functioning mechanisms for
the flow of funds and information.
The same cannot be said of the partnership between the
central government and domestic stakeholders. Higgins
and Rwanyange (2005) present results from a qualitative
assessment of the education reform process in Uganda
focusing on ownership and accountability in the sector.
Based on data collected between 1997 and 2004, the study
finds that officials at the district level acknowledge
improvements in the transfer of funds but feel that they
lack the autonomy to adjust the use of funds to the partic-
ular needs of the districts and that the process is designed
to cater to the needs of the central government and the
donor community.
At the school level, according to the same study, parents
and teachers are excluded from the decision-making process
for the utilization of funds allocated to schools. Some
observers consider this lack of involvement and alienation a
serious constraint to the performance of schools and stu-
dents (Higgins and Rwanyage 2005). In the survey underly-
ing the study, respondents cited the lack of integrity of head
teachers among the factors that impede effective learning
and school management. Some head teachers reportedly fail
to post capitation grants on public notice boards, as
required by the rules.
All stakeholders are concerned about the quality of
entrants to primary teachers’ colleges and the training they
receive there. Teachers themselves feel marginalized,
because they have little say in decisions that affect their
working conditions. These considerations suggest that
domestic partnerships for education in Uganda may not be
as effective as they could be, because of imbalance in influ-
ence and capacity. Education reform in Uganda continues to
evolve, however, and coordination mechanisms have been
improving. Starting in 2007, for instance, the participation
in sector reviews and discussions of the Education Sector
Consultative Committee have been opened up to represen-
tatives of teachers and head teachers, local authorities, and
parent teacher associations.10
CONCLUSION
Uganda’s success in increasing access and equity in primary
education stems from the following institutional factors:
■ Strong political commitment to a development strategy
centered on building human capital
473
 ■ A comprehensive approach to policy making based on care-
ful planning and implementation of critical prior actions
■ Effective and sustained domestic and international part-
nerships supportive of country ownership and donor
cooperation within a sectorwide approach
■ Efficiency gains from measures designed to improve
transparency and accountability at the school level in the
use of available resources.
Reform initially put enormous stress on the country’s
educational infrastructure, reducing the quality of educa-
tion as well as completion rates. Significant improvements
have been achieved over the past few years. Policymakers in
Uganda must now consolidate and expand these improve-
ments if they are to successfully link the country to the
global economy, which pays a premium for knowledge and
skills.

474 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
ANNEX 26A DIFFERENCE-IN-DIFFERENCES
IDENTIFICATION OF POLICY
IMPACT
When repeated cross-section data are available, the impact
of a policy can be identified and estimated using a two-step
procedure known as difference-in-differences (DID) or
double difference. The first stage involves the reflexive (or
before and after) change in average outcomes of the treat-
ment and control groups. This difference eliminates the
selection bias associated with permanent group differences,
leaving intact the bias as a result of the time trend. The bias
that remains after the first step is removed by the difference
in the second stage, in which the change in the average out-
come of the comparison group is subtracted from that
of the treated group.11 Alternatively, the treatment effect
can be obtained by measuring the difference in the change
in average outcomes across groups (that is, comparing
with- and without-treatment groups) before and after
treatment. The DID method can thus be said to be path
independent, because the same result is achieved regardless
of the sequence in which these two differences are com-
puted. DID estimation can be implemented within a
regression framework, which also offers a convenient way
to obtain the relevant estimates and the associated standard
errors. This framework also makes it easier to add more
groups and time periods using dummy variables (Angrist
and Pischke 2009).12
Deininger (2003) applies a regression model that is con-
sistent with the double difference approach to two nation-
ally representative household surveys: the 1992 Uganda
Integrated Household Survey (UIHS) and the 1999 Uganda
National Household Survey (UNHS). Given that UPE took
effect in 1997, observations for 1992 represent the base case;
schooling outcomes observed in 1999 presumably reflect
continued implementation of UPE.
To see clearly what is involved, let Sit stand for an indica-
tor of enrollment status (enrolled or not enrolled) of child i
in year t = 1992 or 1999. Think of this outcome as a func-
tion of individual, household, and community characteris-
tics (Xit) such as gender, income, and parental education.
This indicator also depends on a time trend (T) and unob-
servables (eit) assumed to be independently and identically
distributed. The time variable T is a dummy variable that is
equal to one for 1999 (to mark exposure to UPE regime)
and zero for 1992 (indicating no exposure). Deininger’s
(2003) regression is of the following form:
Sit = a + bXit + g T + dXitT + eit .
CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
The interaction terms in this equation play a crucial role in
the assessment and interpretation of various effects of the shift
to the UPE regime. One can think of the dummy variable T
(representing the shift in policy regime) as having a moderat-
ing influence on the effects of the individual and household
characteristics on the schooling outcome (here the probability
of enrollment). The dummy variable T is a moderator vari-
able.13 The coefficient (g) associated with the moderator vari-
able represents the effect of any secular trend captured by T.
The other coefficients in the model are easily interpreted
within the logic of difference-in-difference estimation.
NOTES
1. The net primary enrollment ratio is the ratio between
the number of children of primary school age enrolled in
primary school and the number of children of primary
school age. By definition, this ratio cannot exceed 100 per-
cent. The gross primary enrollment ratio is the ratio
between the number of children enrolled in primary school
and the number of children of primary school age. It can
exceed 100 percent if children above (or below) primary
school age are in primary school (because, for example, of
repetition or delayed entry).
2. The pupil-teacher ratio rose from 38 pupils per teacher
in 1996 to about 52 in 1997. The situation has improved
since 2000, with the pupil-teacher ratio falling steadily from
65 in 2000 to an estimated 47 in 2010, the pupil-classroom
ratio falling from 106 in 2000 to an estimated 66 in 2010,
and the proportion of untrained teachers declining contin-
uously (MOES 2010).
3. This section relies heavily on earlier work by the author
and two of his colleagues (Essama-Nssah, Leite, and Simler
2008).
4. The 2001 DHS EdData survey collected information on
the age at which children started and finished schooling, edu-
cational attainment, reasons for nonenrollment, the extent to
which parents and guardians were aware of the UPE program.
The survey also contains information about the assessment by
parents and guardians of the quality of local schools.
5. Such methods include central government provision,
contracting out to the private or nongovernmental
sector, decentralization to local governments, community
participation, and direct transfers to households (World
Bank 2003).
6. Parents paid up to 90 percent of recurrent and capital
expenditure for primary education in 1991 (Oketch and
Rolleston 2007).
7. For students in P7, the government also pays the regis-
tration fees for the National Examination Board (Bategeka
2005).
475
 8. For instance, guidelines from the MOES require pri-
mary schools to spend capitation grant as follows: 50 per-
cent on instructional material, 30 percent on co-curricular
activities such as sports and clubs, 15 percent on utilities
and maintenance, and 5 percent on school administration
(Bategeka 2005).
9. According to the commission, primary school comple-
tion is the minimum level of education that all citizens need
to be able to live a full life. It noted that society would ben-
efit a great deal from this outcome, which would probably
increase national unity, moral standards, and prosperity
(MOES 1999).
10. Communication with Luis Benveniste, World Bank, May
24, 2010.
11. The first stage relies on the assumption of time invari-
ance of the group effects; the second is justified by assuming
that, in the absence of treatment, the average outcome of the
treated would follow the same time path as the average out-
come of the control group. When combined with the fact
that both groups are observed before treatment, the
assumption of common time trend makes it possible to
frame the DID method within the logic of the traditional
counterfactual approach. The treatment effect can thus be
identified and measured by comparing the observed out-
come of the treated with the counterfactual predicted on the
basis of the evolution of the average outcome of the control
group.
12. The double difference approach (in which impact is
defined by the difference in the change of outcomes for the
treated and untreated group) can be implemented by a lin-
ear regression that includes a constant, a group dummy
variable, a time dummy, and an interaction term between
the two dummies. Impact is measured by the coefficient
associated with this interaction term.
13. By definition, a moderator is a qualitative or quantita-
tive variable that affects the direction or strength of the rela-
tionship between a response variable and a predictor or
independent variable (Baron and Kennedy 1986). For
instance, the shift in policy regime (the moderator) may
affect the impact of gender (a predictor) on schooling out-
comes (response variables).
REFERENCES
Angrist, Joshua D., and Jörn-Steffen Pischke. 2009. Mostly
Harmless Econometrics: An Empiricist’s Companion.
Princeton, NJ: Princeton University Press.
Appleton, Simon. 2001. “What Can We Expect from Univer-
sal Primary Education?” In Uganda’s Recovery: The Role
of Farms, Firms and Government, ed. Ritva Reinikka and
Paul Collier. Washington, DC: World Bank.
476 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
Avenstrup, Roger. 2006. “Reducing Poverty through Free
Primary Education: Learning from the Experiences of
Kenya, Lesotho, Malawi, and Uganda.” In Attacking
Africa’s Poverty: Experience from the Ground, ed. Louise
Fox and Robert Liebenthal. Washington, DC: World
Bank.
Avenstrup, Roger, Liang Xiaoyan, and Nellemann Soren.
2004. “Kenya, Lesotho, Malawi and Uganda: Universal
Primary Education and Poverty Reduction.” Paper pre-
sented at the conference “Scaling Up Poverty Reduction:
A Global Learning Process and Conference,” Shanghai,
May 25–27.
Baron, Reuben M., and David A. Kennedy. 1986. “The
Moderator-Mediator Variable Distinction in Social Psy-
chological Research: Conceptual, Strategic and Statistical
Considerations.” Journal of Personality and Social Psychol-
ogy 51 (6): 1173–82.
Bategeka, Lawrence. 2005. Universal Primary Education
(UPE) in Uganda: Report to the Inter-Regional Inequality
Facility. Institute of Development Studies, University of
Sussex, UK.
Deininger, Klaus. 2003. “Does Cost of Schooling Affect
Enrollment by the Poor? Universal Primary Education in
Uganda.” Economics of Education Review 22: 291–305.
Devarajan, Shantayanan, and Ritva Reinikka. 2004. “Making
Services Work for Poor People.” Journal of African
Economies 13 (1): i142–i166.
Dixit, Avinash, K. 1996. The Making of Economic Policy: A
Transaction-Cost Politics Perspective. Cambridge, MA:
MIT Press.
Eilor, Joseph. 2004. Education and Sector-Wide Approach in
Uganda. International Institute for Educational Plan-
ning, UNESCO, Paris.
Essama-Nssah, B., Phillippe G. Leite, and Kenneth R. Sim-
ler. 2008. “Achieving Universal Primary and Secondary
Education in Uganda Access and Equity Considerations.”
World Bank, Poverty Reduction and Equity Group,
Washington, DC.
Government of Uganda. 2010. National Development Plan,
2010/11–2014/15. Kampala.
Grogan, Louise. 2006. “Who Benefits from Universal Pri-
mary Education in Uganda.” Department of Economics,
University of Guelph, Ontario.
———. 2009. “Universal Primary Education and School
Entry in Uganda.” Journal of African Economies 18 (2):
183–211.
Heckman, James J. 1976. “The Common Structure of Statis-
tical Models of Truncation, Sample Selection and Lim-
ited Dependent Variables and a Simple Estimator for
such Models.” Annals of Economic and Social Measure-
ment 5 (4): 475–92.
Higgins, Liz, and Rosemary Rwanyange. 2005. “Ownership
in the Education Reform Process in Uganda.” Compare: A
Journal of Comparative and International Education 35
(1): 7–26.
Ingram, Gregory K. 2004. “Overview.” In Evaluation and
Development: The Partnership Dimension, ed. Andres
Liebenthal, Osvaldo N. Feinstein, and Gregory K.
Ingram. New Brunswick, NJ: Transactions Publishers.
Mackinnon, John, and Ritva Reinikka. 2000. “Lessons from
Uganda on Strategies to Fight Poverty.” Policy Research
Working Paper 2440, World Bank, Washington, DC.
MOES (Ministry of Education and Sports, Uganda). 1999.
The Ugandan Experience of Universal Primary Education.
Kampala.
———. 2010. The Education and Sports Annual Performance
Report. ESAPR. for the Financial Year 2008/09. Kampala.
Nelson, Jane, and Simon Zadek. 2000. Partnership Alchemy:
New Social Partnerships in Europe. Copenhagen Centre,
Copenhagen.
Nishimura, Mikiko, Yamano Takashi, and Sasaoka Yuichi.
2008. “Impacts of the Universal Primary Education Pol-
icy on Educational Attainment and Private Costs in
Rural Uganda.” International Journal of Educational
Development 28: 161–75.
Oketch, Moses O., and M. Rolleston Caine M. 2007. “Poli-
cies on Free Primary and Secondary Education in East
Africa: A Review of the Literature.” Research Monograph
CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION
10, Consortium for Research on Educational Access,
Transitions and Equity, Institute of Education, University
of London.
Penny, Alan, Michael Ward, Tony Read, and Hazel Bines.
2008. “Education Sector Reform: The Uganda Experi-
ence.” International Journal of Educational Development
28: 268–85.
Picciotto, Robert. 2004. “The Logic of Partnership.” In
Evaluation and Development: The Partnership Dimen-
sion, ed. Andres Liebenthal, Osvalso N. Feinstein, and
Gregory K. Ingram. New Brunswick, NJ: Transactions
Publishers.
Reinikka, Ritva, and Jakob Svensson. 2004. “Local Capture:
Evidence from a Central Government Transfer Program
in Uganda.” Quarterly Journal of Economics 119 (2):
679–705.
———. 2007. “The Returns from Reducing Corruption:
Evidence from Education in Uganda.” CEPR Discussion
Paper 6363, Center for Economic Policy Research, Wash-
ington, DC.
World Bank. 2003. World Development Report 2004: Making
Services Work for Poor People. Washington, DC: World
Bank.
World Bank, and UNICEF (United Nations Children’s
Fund). 2009. Abolishing School Fees in Africa: Lessons from
Ethiopia, Ghana, Kenya, Malawi and Mozambique. Wash-
ington, DC: World Bank.
477
 Eco-Audit
Environmental Benefits Statement
The World Bank is committed to preserving Saved:
endangered forests and natural resources. The
Office of the Publisher has chosen to print • 12 trees
Yes Africa Can: Success Stories from a Dynamic • 369 pound of solid waste
Continent on recycled paper including 50% • 5,820 gallons of water
post-consumer recycled fiber in accordance • 1,290 lbs Co2 equivalent
with the recommended standards for paper of net greenhouse gases
usage set by the Green Press Initiative, a • 5 million BTU’s of total
nonprofit program supporting publishers in energy
using fiber that is not sourced from endan-
gered forests. For more information, visit
www.greenpressinitiative.org.
 rawing on the existing knowledge of African development from previous

D publications Yes Africa Can takes an in-depth look at 26 economic and

social development successes in Sub-Saharan African countries—20 from
individual countries and 6 that cut across the region. These stories manifest at
the project, provincial, sub-national, national, or regional level and across themes,
programs, and sectors. The book aims to address how Sub-Saharan African
countries have overcome major development challenges. The main components
of each case study include a description of the achievement and the elements
that qualify the outcome as successful; an assessment of the main policies,
interventions, actions, and other factors that contributed to the positive outcome;
a presentation of the lessons learned and the contribution to the discourse
on African development; and insights on the usability or applicability of the
achievement in terms of the potential for scaling up the interventions and actions.
The studies can be classified into four categories: overcoming or avoiding massive
government failure, rebuilding or creating a government, rationalizing government
involvement in markets, and listening to the people.

ISBN 978-0-8213-8745-0

SKU 18745